Final Report Summary - PREDICT (PREparing for the Domino effect in Crisis siTuations)
Executive Summary:
Context
“A small cause may have a big effect!” – Decision-makers in crisis situations and Critical Infrastructure (CI) organisations must consider many factors: political, social, legal, cultural, ethical and economic parameters are always to be considered during threat assessments and counter-measures. The increasing complexity of crisis and disasters have mobilized crisis management organisations and motivated the industry and the scientific community to create tools that will be able to help crisis managers to act more quickly and effectively.
The challenge
The aim of the PREDICT project is to provide an innovative set of tools to deal with cascading effects in multi-sectorial crisis situations covering aspects of critical infrastructures. The PREDICT solution consists of a suite of decision support tools (DST) integrating different services facilitating foresight, prediction, communication and eventually decision making in crisis situation. This suite of tools will be based on improved and innovative methodologies, models and software tools.
Project Objectives:
The main objective is to reduce the impact of crisis with cascading effects by:
■ Increasing awareness and understanding of cascading effects.
■ Enhancing the preparedness of crisis management practitioners for cascading effects.
■ Improving their capabilities to respond to crisis.
Main Results:
The most visible results of the project is:
■ A new advanced decision support and training prototype: the iPDT.
But more outcomes have been obtained from the PREDICT project:
■ 3 test-cases across Europe.
■ A cascading effect assessment methodology.
■ A training methodology.
■ Dependencies & interdependencies information.
■ A solid pan-European network of crisis management practitioners.
Project Context and Objectives:
Context
The aim of the PREDICT project was to provide a comprehensive solution for dealing with cascading effects in multi-sectoral crisis situations covering aspects of critical infrastructures.
The PREDICT integrated Tool (iPDT) consists of a suite of decision support tools (DST) integrating different services facilitating foresight, prediction, communication and ultimately decision making in crisis.
This integrated tool is based on improved and innovative methodologies, models and software tools. It aims to increase the awareness and understanding of cascading effects by crisis response organisations, enhance their preparedness and improve their capability to act in case of cascading failures.
Definitions
To use the same terminology and have the same understanding, the PREDICT project selected the following definitions:
Cascading failures: “A cascading failure occurs when a disruption in one infrastructure causes the failure of a component in a second infrastructure, which subsequently causes a disruption in the second infrastructure.” (Rinaldi, 2001)
Critical Infrastructures: “Critical infrastructure’ means an asset, system or part thereof located in Member States which is essential for the maintenance of vital societal functions, health, safety, security, economic or social well-being of people, and the disruption or destruction of which would have a significant impact in a Member State because of the failure to maintain those functions.” (European Council Directive 2008/114/EC)
Objectives
The key objectives of PREDICT were to:
■ Gather and analyse available domain knowledge;
■ Develop a common framework;
■ Create conceptual and executable models of cascading effects and interdependencies;
■ Develop a suite of software tools;
■ Validate the solution through running simulations in the framework of three test cases;
■ Disseminate project results and build appropriate liaisons among stakeholders.
PREDICT project phases
Figure 1 illustrates the different phases of the project to reach our objectives. The first phase of the project consisted in gathering information and performing different analysis to feed the tools’ and models’ developers. This occurred mostly in WP2 & WP3 and in addition, first information from external end-users was gathered in WP8. The second and longest phase of the project was dedicated to the development of the iPDT architecture and of the different models and functionalities. In parallel, the PREDICT test-case were organised by WP7 with a first round of assemblies and two other end-user workshops were also organised with external end-users (WP8). The last phase of the project is the consolidation phase. The iPDT prototype was integrated step by step by clusters. Methodologies were prepared so that the prototype could be tested and discussed during the three PREDICT final test-cases. During the last two end-user workshops, more recent version of iPDT could be presented and tested.
Key results
The 3 main results of the project are as follows:
1. The PREDICT incident evolution framework, which provides a threat quantification methodology allowing for the assessment of cascading effects and the modelling of interdependencies between CIs.
2. Improved versions of tools towards the same objective:
▪ SBR (scenario reasoning): a tool supporting the generation and analysis of most probable set of scenarios;
▪ PROCeed (scenario player): which enables crisis managers to model and simulate crisis scenarios;
▪ Myriad (impacts evaluator): a tool supporting decision by risk-based assessment of the current and predicted situation.
3. The Integrated PREDICT Tool Suite (iPDT) (see Figure 2), which successfully combined these tools to provide a solution enabling decision-makers to generate, run and analyse alternative scenarios of a given crisis, identify crucial dependencies between CIs and act with better understanding of the future.
Furthermore, PREDICT project has delivered a unique set of models and modelling approaches (including objects’ classes, attributes, characteristics and dependencies), which proved to be applicable in three different scenarios (Flooding, Railway Incident and Maritime Incident).
Throughout the PREDICT project, crisis management practitioners have thus followed the progress of work, exchanged best practices with their peers and assisted in defining common needs pertaining to the management of cascading effects. This method of “co-development” ensured that the solutions developed in the project respond to the requirements and constraints of these professionals.
In addition to this, several dissemination activities have been conducted and papers have been published over the course of the project, ensuring that the results of PREDICT are usable and exploitable by the research community.
Project Results:
Project organisation
The PREDICT project was organised in such a way that end-users were deeply involved. They are part of the project at three levels:
■ Three end-user organisations are partner of the PREDICT consortium;
■ End-users member of PREDICT’s Advisory Board ; and
■ Many representatives from end-user organisations across Europe.
Figure 3 illustrates the continuous involvement of end-users in the project activities.
Returning to the work package way of distributing work and responsibilities: in WP7, 3 use cases were developed under the lead of our field partners (UIC – International Union of Railways; VRZHZ – Safety Region South Holland South (NL); and SYKE – Finnish Environment Institute). During the project’s lifetime each year and for each individual use case (scenario) workshops with respective (local) stakeholders have been organized.
■ The workshops in 2014 were beneficial to orient the research activities and to tune the requirements and specifications for the integrated PREDICT tool. The research activities were mostly covered in WP2 (Domain analysis and Requirements) and WP3 (Incident Evolution Framework). The tool development activities started progressively for the backbone of the system that eventually had to integrate the functionalities (WP4) and later for the foresight and prediction functionalities (WP5) and for the decision support functionalities (WP6).
■ The workshops in 2015 and 2016 were held to test and validate advanced mock-ups of individual tools (2015) and the first version of the iPDT integrating the individual tools (2016). The integration of the functionalities was done in clusters and the tool developers worked closely with one of the use case leaders. The functionalities were then not equally developed for the three use cases, but an overall assessment could be done.
Secondly, an end-user network was built in WP8. Five workshops were organized to trigger interest for PREDICT’s objectives and developments and to get further recommendations for iPDT functionalities and applications. The end-user network consisted of a broad set of public crisis managers and infrastructure operators from across Europe. The two final end-user workshops within WP8 allowed the PREDICT partners to discuss the results of the iPDT with different types of end-users and to collect further recommendations for a follow-up of the activities. Next sections elaborate on the setup and context of both approaches, the use cases with end-users, and the end-user network development activities.
WP2 – led by CEA "Domain analysis and Requirements"
Under the title “Domain analysis and Requirements”, WP2 was the baseline activity for the understanding of “Cascading Effects & Resilience” in complex systems through the realisation of a comprehensive assessment. It is a primary step towards the development of PREDICT crisis management methodology and its corresponding integrated DSS. The integrated PREDICT Tool is an end-user oriented tool and it is called “iPDT”.
The main objective of WP2 is to provide a common understanding for the consortium in order to develop a methodology to assess the “cascading effect & resilience” of complex systems. The methodology should admit a given level of interconnectivity and well-defined thresholds. The definition of cascading patterns and the selection of the sets of critical infrastructures will be guided by real crises occurred in the past.
The WP2 was structured in 3 tasks each has its own objective:
■ T2.1; to assess the “state-of-the-art of the R&D in cascade effect & resilience and global modelling”,
■ T2.2; to assess the “security metrics for threats, for systems’ resilience MS&A activities”, and
■ T2.3 to assess the “DSS and predictive tools feature specifications”.
The outputs of WP2 (all three tasks included) are underlying inputs to all other PREDICT’s WPS.
T2.1: State of the art of the R&D activities in Cascade Effect & Resilience and global modelling
The main objective of the task T2.1 is to assess the state-of-the-art on the R&D activities in Cascade Effect & Resilience in crisis situation. The assessment laid down the foundation for a common understanding of the existing views and models in cascade effects and resilience in complex systems. Besides, the assessment has allowed us to determine the complexity of the subject and to evaluate the gap separating between the models and the real end-users’ needs.
The task T2.1 starting point was an extensive survey of the existing approaches of Modelling, Simulation & Analysis (MS&A) of the cascading failures and the CI’s resilience. The survey considered: academic papers; experts reports and letters, books, industrial technical reports, crisis analysis reports and safety/security authorities documents. Out of some 250 different files covering about 10 recent years, at the end 122 have been considered in the assessment because of their relevance. We recall that out target in methodologies, models and tools in cascade effects & resilience of complex systems. A distinction between qualitative and quantitative approaches has been used as an a priori filter.
The state of the art was categorised in main areas that we deemed relevant for PREDICT tasks: threats identification and classification, dependencies and interdependencies between critical infrastructures, crisis management and decision support system (DSS) specifications. Some sub-categories have equally been underlined: failure data and databases identification, failures propagation, threats detecting and periodic monitoring, cascade dynamic modelling and resilience metrics. In addition, D2.1 put some light on the state of the art in taxonomy corresponding to threat, critical infrastructure, and crisis codification.
The assessment covered three macros areas. The 1st macro area is related to qualitative activities which will allow us identifying and classifying the most plausible threats considering past crisis situations. The 2nd macro area is related to the identifying, assessing and commenting the existing quantitative methodologies and models in cascading assessment. The 3rd macro area is dedicated to the development of a taxonomy for threats based on collected past and the running R&D effort and the harmonization of existing taxonomies relative to the global security. The assessment considered the crisis management end-users practices and requirements, as well.
T2.2: Security Metrics for threat and system’s resilience MS&A
Generally, threat identification and characterisation is a first act in almost all MS&A of cascade effects. A pertinent identification and characterisation of threats would necessarily be based on the use of the most appropriate security metrics. The starting point for the development of appropriate security metrics was a comprehensive state-of-art analysis of different existing metrics in different disciplines. This allows a better determination the classes of causes, consequences and sequences.
Task T2.2 aims at addressing those criteria of appropriate security metrics from threat identification & characterisation standpoint. The demarche was based on recalling the current learnt-lessons in some representative fields where some systems (and systems of systems) had been exposed to severe threats (sometimes catastrophic) and the threats had to be managed. Some learnt-lessons could successfully be translated in codes, norms and regulations and subsequently received worldwide recognitions. Others are still in a pre-maturation phase with limited local recognition. Few others are kept at the stage of local good practices with no external visibility.
Equal efforts were put is the task to cover the two main categories of threats: Natural and man-made ones. Only threats with adaptability-features have been excluded: terrorist actions, sabotages and wars.
The deliverable addressed the generic criteria and specifications of the security metrics that could be considered as appropriate for threats identification & specification.
T2.3 DSS and predictive tools feature specifications
Task 2.3 is the “tools feature specifications”. The final specifications of the iPDT will be set in the technical work-packages WP4-5-6, sticking as much as possible to the proposed features recommended by T2.3. The feedback development experience from the other WP’s will be considered to update these recommendations-specifications if necessary. The updating is a dynamic process in order to maintain and to guarantee a common understanding of the cascade effect MS&A, throughout the whole life of the project. The specifications have been critically assessed by WP4 “System Design & Iterative Integration”, as WP4 is to develop and implement the architecture of the iPDT.
WP2 has equally associated end-users in the critical assessment of its recommendations & specifications through a workshop on “MS&A of cascade effects-WP2 findings” (WP8). The end-user expressions & needs for the iPDT have been collected within the consortium and processed. During this workshop, various end-users in all CIP fields provided their inputs. The results of the workshop has been analysed by the project’s tool provider and the system’s architect (THALES, TRT-NL, iTTi, VTT and Fraunhofer). Furthermore, the scenario methodology task, which overlaps with T2.2 has been completed within T2.3.
Task 2.3 was structured in 4 complementary topics.
The first gave a general overview on “Decision Support Systems and Predictive tools”, based on D2.1 assessment outputs. This put the focus on the major feature specifications.
In the second topic, T2.3 assessed the end-users requirements and specifications, from deliverable D8.1 [PREDICTD81] and subsequently updated its recommendations & specifications.
In the third topic, WP2-T2.3 focused on the predictive capability of iDPT.
The last topic is dedicated to the feature specification of “CIPRNet methodology to create crisis scenarios”. Within this topic, examples of scenarios in different crisis situations were treated. These crisis situations covered: flooding, train transporting hazardous substance derailment, Nordic winter storm, coast submersion, earthquake and large forest fire.
Outputs to other PREDICT-WPs
PREDICT-WP2 provides a baseline for understanding cascade effects and resilience in Critical Infrastructures facing given threats. This comprehensive assessment of cascading effects and resilience of systems is a primary step towards the development of the appropriate PREDICT cascade effect Modelling, Simulation and Analysis (MS&A) PREDICT predictive tool suite “iPDT”. WP2 came up with a set of conclusions and recommendations to guide the design of PREDICT Tool Suite for cascade effect MS&A. WP2’s major outputs to other PREDICT WPs were listed in a table in the deliverables and project periodic report.
WP3 – led by VTT "Incident evolution framework"
The main S&T results/foregrounds of WP3 “Incident evolution framework” are the seven-step methodology for assessing cascading effects, the cascade probability function providing a methodology to assess the likelihood of cascading effects, the modelling approach for threat quantification revealing cascading effects of interdependent critical infrastructures (CIs), and further development of the Stochastic Operation Time Model (SOTM) estimating the uncertainty related to human factors.
Seven-step methodology was developed for assessing cascading effects. It describes the steps that a crisis management organisation should perform in order to identify the cascading effects for a specific threat scenario. The steps include 1) identifying the threats to be considered, 2) identifying the CIs in the region, 3) identifying the key CI elements, 4) characterising the vulnerability of the key CI elements to the threat, 5) assessing the first order impact of the threat on the CI elements, 6) describing the dependencies between the CI elements in the region (i.e. describing the required input and output of all key CI elements, distinguishing between the different modes of operation, and including the temporal and spatial factors), and 7) assessing the CI cascading effects. The seven-step methodology can be used by end-user experts when figuring out crisis situation scenarios and providing input data and estimates.
Cascade probability function is a mathematical formulation of the seven-step methodology, showing how the steps can be performed using time-dependent, stochastic input. The modelling framework for the cascading effects of interdependent CIs integrates a threat function which reflects the time evolution of the threat scenario, a vulnerability function which describes the vulnerability of the CI elements for a specific threat intensity, and a dependency function which describes the dependency between the different CI elements in the specific area of consideration. This procedure results in a cascade probability function, providing a methodology to assess the likelihood of cascading effects.
The seven-step methodology is used for providing input data and estimates, and the cascade probability function then generates the final estimates for decision makers. Thus, the seven-step methodology can be seen as “a user interface” and the cascade probability function as “an engine” for calculations.
Modelling approach for threat quantification was developed using the PREDICT case studies as development environment. In the modelling approach, an accident scenario map, locating the initiating events and CIs, is defined and a hexagonal grid is laid on the map. The relevant CIs for each hex are identified, and their interdependencies and vulnerabilities are defined in the model. A reference point is chosen for each hex. The threat function results (provided e.g. by a separate modelling tool) on the reference point are then applied to all CIs in the hex. The initial failure times (i.e. not considering the interdependencies) of the CIs are determined. The final failure times are determined taking into account the interdependencies between the CIs, and the cascading effects are assessed. Thus, the modelling reveals the cascading failures in which a CI is lost due to a failure of another CI.
The modelling approach for threat quantification requires that sufficient data on relevant CIs and their locations, interdependencies, vulnerabilities and initial failure times due to threat functions is available. Depending on the input data available, model parameters can be implemented either as single-value point estimates or as probabilistic distributions. The results can be presented on a timeline to illustrate the incident evolution.
Threat quantification modelling can be utilised in crisis management in both the preparedness and training phase and the response phase. In the preparedness and training phase, the modelling illustrates the progress, influencing factors and potential cascading effects of accident scenarios. It gives guidance for the planning of emergency response by revealing the CIs that are important to protect in order to mitigate or prevent the escalation of the accident. To support contingency planning, the threat quantification modelling can reveal crucial vulnerabilities and dependencies which should be eliminated or mitigated to strengthen the resilience of the CIs. In the response phase, scenarios pre-examined in the preparedness phase can be used as references to support decision making. New simulation results on the threatening phenomena can be input to the threat quantification model to correspond the real-life accident, and predefined vulnerabilities and dependencies can be adjusted by expert judgement using the information obtained from the evolution of the crisis.
Stochastic Operation Time Model (SOTM) was originally developed under the Finnish national research programme for nuclear power plant safety. This model has been further developed in PREDICT WP3 to take into account the uncertainty associated to human actions, and the consequences of delays which cause failure risks in response operations. The novelty of the methodology is related to the combination of temporal and reliability aspects: failures in human operations are not considered as final stages but rather as additional time delays of a specific action. Furthermore, the development work aimed at making SOTM feasible for and implementable to iPDT, the integrated PREDICT decision support tool.
SOTM is based on the assumption that human operations can be described as time delays and possible additional delays due to unexpected factors. In WP3, a mathematical description of the model was presented and a generalised methodology for the construction of the model was developed. The methodology consists of modelling steps for the stochastic estimation of operation times, and timeline charts to visualise the action and communication processes of response procedures. Carrying out Monte Carlo analysis, a probability distribution for the total duration of response operations is obtained. By comparison to the temporal development of the crisis, it can be assessed with which probability the response actions can prevent the escalation of the situation. The modelling gives insight into interdependencies of various phenomena and activities of different actors, and can reveal bottlenecks and contradictions in processes and organisations.
A set of methodologies for the specification of model parameters related to human operations was studied, and the use of the concept of performance-shaping factors (PSFs) was suggested. Based on the applicability and availability of the modelling options in the iPDT development suite, agent modelling was chosen for further investigation and for implementation.
The seven-step methodology, the cascade probability function (including threat, vulnerability and dependency functions), the threat quantification methodology with its various phases, and the stochastic operation time model with the set of methodologies for the specification of model structures and the parameters related to human operations have been forwarded to WP5 and WP6 for implementation into iPDT.
WP4 – led by Fraunhofer “System Design & Iterative Integration"
The main objective of WP4 was to develop a software architecture for integrating FPT (developed in WP5) and DST (WP6) into one coherent technical solution for cascading effects analysis in civil domains. Based on this architecture, the iterative system integration process needed to be conducted by WP4 to achieve the final integrated system. These two objectives have been successfully achieved by WP4 (together with other WPs) in PREDICT. The main results/foregrounds are:
• Software architecture for system integration based on the MCRI concept. WP4 proposed an architectural approach for integrating information systems used in a typical crisis management process, which consists of three parts: situational awareness, foresight and prediction, decision making. This architecture tries to combine modern software engineering best practices with the specific requirements in crisis management process. RESTful mock-up services based on modern Service-Oriented Architecture (SOA) are the essential parts of the integration architecture. Reverse- proxy based solution provides a flexible runtime environment for hiding the technical details of different system implementations. Special design consideration is also given to integrate spatial data into the crisis management process. To maximise the system design flexibility, software containers are used to provide flexible wrappers for the real implementation. This architecture is proven to be effective for the integration work in PREDICT.
• The iterative integration approach. Together with the integration architecture, the iterative integration approach is also one of the major outputs of WP4. Working with partners from different organisations on the same software project can be difficult, especially when it comes to integrating new system features and providing system maintenance. It can yield unwanted dependencies and slow down the software development process. An iterative approach of system integration can be separated into three stages:
1. Defining specification and requirement of the service. This includes developing use cases, formal specification, etc.
2. Writing service mock-ups and deploy them to the server for automated testing. After this stage, all unit tests should pass as required in classical Test-Driven Development (TDD).
3. Iteratively replace mock-ups by real implementations. Each time, if a service mock-up is replaced, all unit tests must be executed to guarantee that the service implementation meets the requirements defined in the specification.
The first step required a complete specification description of each RESTful endpoint including:
1. Uniform Resource Locator (URL) - a unique ID of the service.
2. Request method type - indicating the character of the service whether it is for reading, writing or deleting operations.
3. Header information - some meta information that is not suitable to be encoded in the URL.
4. Payload - additional information that is two large to be encoded in the URL.
5. Expected responses - the result delivered by the service implementation.
Finally, after all the mock-ups were replaced by real implementation, it is always a good practice to have some high-level tests running like functional tests and even human-guided tests.
• The integrated system iPDT for cascading effects analysis. iPDT is the major technical output of WP4 with the tools developed in WP5 (PROCeed) and WP6 (MYRIAD and SBR). Regarding the result of the integration, iPDT is characterised as a system that delivers two ranges of functionality. First, it delivers system-level user functions for accessing iPDT and its component systems (like a user login) and a developer access for monitoring and maintenance. Second, the main added-value of iPDT is that it enables its component systems to use the other components as additional services to provide integrated functionality. There is no specific iPDT GUI, but, for instance, in “iPDT mode” the PROCeed GUI provides more functions than in “stand-alone mode” and can display information generated by the other system components MYRIAD and SBR. iPDT is implemented in a distributed fashion. The physical core of iPDT is a reverse proxy server hosted at Fraunhofer. It connects to the partner’s servers that run the other system components (FPT, DST, middleware, external tools). The integrated system has been demonstrated at three end-user workshops organised by WP7. Feedbacks coming from the end-users during these workshops provide invaluable information for the future improvement of iPDT.
WP5 – led by iTTi "WP5 “Foresight and Prediction Tools”
There were two main objectives of the WP5:
1. Design and develop of the Foresight and Prediction component for the iPDT solution ; and
2. Design and develop appropriate interfaces for the WP5 and WP6 tools communication
Both objectives has been fully achieved. The most important technical results are:
a. Simulation Engine Development
The concept of the PREDICT – FPT simulation was to follow a stateless approach, which is partially based on the Discrete – Event Simulation Concept. Thus there is no continuous simulation being processed, as there is a simulation on demand (step by step approach) implemented. Hence the interactions (also referred as ‘actions’) with the user and external tools are required. Following interactions are distinguished in the PREDICT – FPT simulation: (i) sending and receiving simulation data from/to external tool, (ii) ad-hoc interactions with the user or user’s action, and (iii) execution of a simulation module of the PREDICT – FPT.
The simulation in the PREDICT project is a process of interpreting and processing chronologically ordered events and calculating the future course of events provided by the simulation engine and supported by other PREDICT functionalities (e.g. SBR) and simulation modules (e.g. rule engine).
b. Rules Engine Development
In order to effectively simulate the evolution of the crisis, a Rule Engine has been developed as part of the Simulation Engine. The rules on the conceptual side are based on Drools (www.drools.org - a business rule management system) which is a well-known solution dedicated for a business process management. The syntax of the rules is based on MVEL (MVFLEX Expression Language) – a hybrid dynamic/statically typed, embeddable Expression Language and runtime for the Java Platform. The rules on the conceptual side are based on Drools business rule management system, which is a well-known solution dedicated for a business process management. The syntax of the rules is based on mvel2 (MVFLEX Expression Language) – a hybrid dynamic/statically typed, embeddable Expression Language and runtime for the Java Platform.
c. Development of PREDICT-FPT Modules
The PREDICT-FPT component follows the concept of object-oriented modelling and object-oriented design. Consequently, data storing follows a concept of object-relational mapping. Development of the PREDICT-FPT component was based on the following modules: (i) Users' Actions Module, (ii) Object Management Modules, (iii) Rules and Interactions Module and (iv) Scenario Management Modules. Each module is controlled by the Simulation Engine which is interfaced with the PROCeed tool and generates simulation results. The generated results are available for the iPDT, which is the main data broker available for all the modules and tools running under PREDICT–FPT and PREDICT–DST subsystems.
d. Development of Interfaces for Decision Support Tools
The PREDICT-FPT component is responsible for managing the simulation (scenario evaluation) process. The communication between different components is based on RESTful APIs of the tools. The communication between the tools is synchronous and the initiator of the communication is the PREDICT-FPT component. Exceptionally other data sources treated as sensors or user interaction are based on the asynchronous communication. It is of crucial importance that a single event describes the change of an object and each message (describing an event) consists of the object identifier, time of the event occurrence and event type (i.e. predefined; created by user; generated by external tool; detected by sensor). The stateless simulation approach is used within the data integration.
e. Technical implementation of the WP3 Incident Evolution Framework
WP3 provided a detailed description of the 7-step methodology for the cascading effects identification and assessment. Several recommendations for the technical implementation were provided as well. PREDICT-FPT followed the methodology itself and provided recommendations. Thus crucial elements of the cascading effects identification process were fully implemented in the software. Thus the user is enabled to conduct set of actions within the tool, which lead to cascading effects identification and assessment. Following activities are considered to be crucial from the cascading effects identification perspective:
(i) Identification of the threats to be considered at the particular location;
(ii) Identification of the Critical Infrastructure located in the region;
(iii) Identification of the key CI elements;
(iv) Implementation of the vulnerability function and first order impact assessment;
The interconnections between the 5 modules developed is illustrated in Figure 5.
WP6 – led by THALES “Decision Support Tools”
The objective of this WP was to predict the range of possible consequences of each decision, and to determine, as accurately as possible, the probability that each credible consequence will occur.
WP6 has fully achieved his goals. The main results are:
• Establish formal and mathematical models of risk indicators
• Establish a hierarchical task analysis (HTA) describing the goals and tasks involved in the decision making processes
• Multi Criteria Decision Making approach, based on a methodology that models and solves the MCDM problems with imperfect information
• A tool proposing a list of actions mitigating the risks to achieve the goals of the expert
• A Dynamic Expert Integration Network that supports a systematic combination of fusion and analysis services into complex assessment systems with humans in the loop
• A tool that facilitate description of uncertainties by developing alternative scenarios
• These achievements are documented in deliverables and demonstrated through the tools in the iPDT suite.
The results of each task is summarized in the following paragraphs:
Task 6.1 Risk Indicators Requirement and Specifications
The deliverable D6.1 summarizing the work done during this task. It states the “Requirements and specifications report of Decision Support tools” provides a structured framework to construct an aggregated risk, fit for use in cascading effect situations. In a more detailed way, this deliverable presents, first, a method: the Multi-Criteria Decision Making Approach (MCDA), to identify the risks indicators requirements and specifications. Next, this document presents the hierarchy of point of view based on the three components of risk assessment for cascading effects: likelihood, impact and topology of the network and defines theses points of views thanks to the key relevant questions answered by the end-users. The document ends by describing how we can operationalize the points of views with metrics and aggregations adapted to the Predict needs with a mathematical structure allowing formalizing and defining of “non-naturally numerical” impacts, which can then be used to contribute to the risk level, and focused on criticality and geography of the affected areas, as well as impact of time.
Task 6.2 Cognitive analysis to support decision tools specifications
In order to do the cognitive task analysis a simulation of a ROT (regional operational team) situation. A working session was organised during a test-case workshop on October 15th, 2015 in Dordrecht. The participants of the ROT simulation were asked to consider the flooding case and the possibility of an evacuation. During this simulation we made an initial attempt to do a hierarchical task analysis (HTA) describing the goals and tasks involved in the decision making processes. Due to the dynamics of the discussion, it was difficult to obtain the HTA during this event. The simulation was recorded on video for further post analysis. Based on the video material another workshop was organized on November, 26th in Delft to analyse results of video material & transcripts. Additionally, historical accident reports related to railways were analysed. We were able to specify an HTA tree, test it for the three test-cases and identify where the PREDICT tools could be used in the decision making process of ROT members. The details of this research are available in Deliverable D6.2.
Task 6.3 Multi-criteria decision making under uncertainty
This task is the topic of D6.3. It describes the technical aspects of the risk assessment module which is based on multi-criteria decision making under uncertainty. This is the core of the “impacts evaluator” of the iPDT. It provides an analysis to choose a multi-criteria decision making approach by identifying the functional requirements that shape the module, and derives a suitable MCDM model and its functionalities. It also tackles the definition and handling of uncertainty in the module: imprecise values of the input, of the parameters of the decision model (e.g. weights, value functions) itself, partial relevance of the sectors in each use-case. The multi-criteria decision model shall thus be reshaped automatically to focus only on the relevant sectors, using for instance expected utility, pessimistic rule or min-max regret.
Task 6.4 Uncertain decision making tool to support short or mid-term planning
Deliverable D6.4 summarizes the approaches developed to assess the impact and risk of cascading effects, which have been implemented in the MYRIAD tool. This includes conversion of scenario trees into individual sector assessments, and identification of optimal sequence of actions (i.e. decisions). Three methods are proposed: identification of the “most probable” scenario (adapted to PROCeed workflow); computation of the expected risk given uncertainties; computation of the min-max regret (minimization of the regret). The implementation of the MYRIAD tool is also described in D6.4 and shows how the user can modify the preference information to adapt to a new use-case.
Task 6.5 Dynamic Expert Integration Network
The Dynamic Expert Integration Network (DEIN) is an information sharing tool for experts to quickly share critical information during large scale crisis events. This tool which is not fully integrated to iPDT is more operational oriented. DEIN is a special instance of the Dynamic Process Integration Framework (DPIF) that focusses only on integration of human expert instead of automated processes. Since DPIF is a SOA the communication in DEIN is service-based. This means that all communications between different experts is facilitated through the expert’s provided and required services. In that sense, it is very different from other type of communication systems, such as the telephone, e-mail, chat, etc., where you have to know the contact details of the person you want to communicate with. In DEIN communication is established based on the type of information you, as an expert, can provide or need. The person (typically another expert) that needs or can provide critical information does not necessarily have to be known in order to communicate since the communication is service-based. For example, a service can be provided by an organization with multiple experts that can actually provide this service. In this way redundancy is automatically build in and different experts can provide time sensitive information quickly.
Task 6.6 Scenario-based Reasoning Tool
The Scenario-based reasoning (SBR) tool consists of a “frontend” and a “backend” part (see Figure 7). The frontend solely consists of a browser-based GUI that uses RESTful web services to communicate with the backend (see the communication channel depicted as arrows in Figure 7).
With the help of SBR frontend the end-users can view different aspects of CIs (like states, location, etc.) and other type of objects like schools, hospitals etc. that are involved in different scenarios for investigation. For each CI/object in SBR, the user can view the likelihood of CI/object failure at a certain point in time. Additionally, the graphical user interface supports the visualization to up to three alternative developments (most likely, worst-case and best-case scenarios) for a given scenario concerning the different failure times of CI. These possible scenarios are directly presented in the timeline component (see the bottom part of Figure 8 annotated with an orange rectangle) of the GUI. By clicking on the red or green icons shown in the timeline (see Figure 8), the user can easily find the corresponding CI/object on the map view.
Different scenarios can be inspected by setting observation about the CI/objects shown in the map (see the green block in Figure 8). For example the user can tell the system that a certain CI failed at a certain time by setting the value of CI state to not working. Based on this observation, the SBR engine is able to compute the likelihoods of failure for the other CIs/objects by considering the dependency model embedded in the test case. This allows the user to inspect the failure time of different CI/objects under different conditions (what-if exploration).
WP7 – led by TNO “Prototype and Testing Validation”
The purpose of Work Package 7 (WP7) within the PREDICT project is to appreciate (elements of) the iPDT with end-users who are acting in the three PREDICT test-cases. For that matter, a series of engagements with end-users has been organised to identify their requirements, to present first versions of elements of the iPDT, and to get an appreciation of the iPDT in its final stage with recommendations from end-users for the way ahead. The specific use cases were developed under the lead of our field partners (UIC – International Union of Railways; VRZHZ – Safety Region South Holland South (NL); and SYKE – Finnish Environment Institute).
During the project’s lifetime each year and for each individual use case workshops with local stakeholders have been organized.
■ The workshops in 2014 were beneficial to orient the research activities and to tune the requirements and specifications for the integrated PREDICT tool.
■ The workshops in 2015 and 2016 were held to test and validate advanced mock-ups of individual tools (2015) and the first version of the iPDT integrating the individual tools (2016).
The WP7 S&T outcome can be divided into three parts:
■ General approach → for testing the PREDICT results
■ Validation results → through case studies
■ Training methodology → to support end-users in developing a training using PREDICT results
General Approach
A framework for prototype testing in case studies allowing for the evaluation of the applicability and added value by the PREDICT tools and methods suite has been developed. It was decided to have the end-user feedback on the tools in a two-step approach. In the first step concepts of the tools in mock-up versions will be presented to the end-users. With the feedback then received, the tools have been further developed into prototype versions and integrated in the PREDICT tool. This gives the maximum opportunity for end-user to influence the development of the tools. The final workshops then have the purpose of assessing the value added of the iPDT in the crisis management process.
A set of evaluation criteria has been offered based on which assessment of the added value, both from operational and organizational perspective. To assess the iPDT potential the following criteria were taken into account:
Functional: does the tool provide the functions required for managing cascading effects?
Information: does the tool provide the information required for managing cascading effects?
Organizational fit: does the tool fit the organization of end-users?
Overall acceptance: do end-users intend to use this technology?
Evaluation schemes and forms have been prepared for a uniform evaluations by individual participants to workshops and for condensed workshop reporting.
Validation results
Two series of workshops were planned. For the first workshop series the tool developers have, in close alignment with case leaders, prepared a mock-up for their tools for demonstration in the selected scenarios. The conclusions of these workshops can be summarized as follows:
■ The tool(s) have training value for the operators, and are considered useful for planning purposes;
■ For the operational phase (hot phase of the crisis), end-users don’t see a need for additional tools;
■ The possibility to run different crisis situation scenarios was much appreciated by the users;
■ Regular meetings with the case leader and the different stakeholders are essential to revisit the value added of tool and its results (usefulness, reliability);
■ Individual quality and value of tools should have the highest priority; and only then to seek for integration in a suite of tools;
■ It is important for the end-users to know what kind of models are implemented in the tool.
Referring to the above conclusions it has been decided to not use iPDT in an operational setting during the PREDICT project timeframe. The second series of workshops therefore was developed for a training setting and to use the iPDT as supporting tool for this training. During the workshop in Dordrecht a PREDICT team member supported the training of the responsible crisis management team with cascading effect information based on iPDT. At the Paris workshop (UIC), the iPDT has been demonstrated to stakeholders from a far broader community than only crisis response organisations. At the Helsinki workshop a training regarding cascading effects has been planned at which the iPDT has been demonstrated and a hands-on session has been held. All workshops offered the possibility to extensively discuss the usefulness and the necessary add-ons needed for future use in practice.
The overall results of the second workshop series confirmed the earlier outcome that end-users appreciate the iPDT rather for planning and training purposes, with little or no use for preparation in case of an emerging crisis or during a crisis. They could foresee that the IPDT may work in their organization. This would however require adjustment of the procedures for planning and training and optimization of the iPDT for this use.
Users positively appreciate the functions and information in support of creating awareness of failing critical infrastructures and, to a lesser degree, for understanding cascading effects. Since decision support (MYRIAD component) was not yet fully integrated, end-users were not able to appreciate support for making decisions and missed information about the effects of alternative courses of action.
The discussions provided good insights for improving iPDT. The system has potential to try out scenarios of cascading effects and to get insight in the effects of alternative courses of actions. End-users like to use the iPDT for training “what-if” scenarios; to learn about short and long term consequences; to compare alternative courses of action; to learn about the impact on society and whom to inform and coordinate with. To realize this potential the following is recommended:
■ Integrate and assess the modules for supporting decisions about alternative courses of action.
■ Optimize iPDT for training
■ Optimize iPDT for planning (risk analysis and management)
■ Identify stakeholders and end-users to test new functions for the planning and training use case
■ Develop guidance for using iPDT for training and planning purposes
■ Develop databases with scenarios for training and planning
■ Use recommendations from end-users to develop iPDT for training
Finally, participants at the workshops underlined their interest in hands-on experience with the iPDT in assessments of next versions of the iPDT after the project.
Training methodology
A generic approach for training methodology is based on training requirements, supporting training tools, the actual training and evaluation of training results. This can be applied to various levels of trainees (operators, technicians, trainers and even tool developers).
A stepwise methodology (Event Based Approach to Training - EBAT) have been provided to develop such necessary training and examples showed how this method can be used for training of crisis managers using iPDT. The most important recommendations from the EBAT approach are:
■ To develop learning experiences or learning tasks based on learning goals and not on operational tasks
■ To specify and introduce specific events in any training session that elicit the behaviour that contribute to the acquisition of necessary knowledge and skills
■ To create a meaningful learning environment for trainees where they can experience consequences of correct and incorrect behaviour, and receive feedback on how to improve their behaviour.
The EBAT approach has been used to prepare one of the workshops in the second series. Practical guidelines to develop exercises using a structured approach (including templates) have been proposed and used. This structured approach will support the exercise developer and tool experts to work together to develop an exercise to gain awareness, understand and manage cascading effects supported by the capabilities offered by iPDT. The possibilities of iPDT in combination with this structured approach as a tool for training are evaluated:
■ For iPDT to be used for other scenarios a more extensive validation exercise is necessary for a fully developed iPDT
■ Pre-defined learning objectives can be reached when using the offered structured approach supported by templates
■ The iPDT evokes a learning process: decisions and actions must be made and it is possible to simulate the effect of different decisions in the tool.
■ During exercises supported with iPDT, evaluators should be used and instructed to provide feedback.
■ Actual integration in the iPDT with connectivity to already operational systems leads to consistency over multiple systems
Based on the results, it can be concluded that iPDT is suitable as a tool for training, especially when used in combination with the templates and dedicated evaluators to provide feedback. iPDT in combination with the structured EBAT approach to exercise development, supported by the templates, forms a strong ensemble for training purposes.
The end-users foresee that the iPDT may work in their organization for training or planning. Adoption would be facilitated by a clear business case for the iPDT in which the benefits and costs of using the iPDT are formulated and supported with evidence. Adoption would also be facilitated by procedures for planning and training that are compatible with an improved version of the iPDT.
Optimization for planning requires specification of risk analysis and management in planning processes (e.g. goals, activities, results in existing procedures) and identification of requirements of end-users for using the iPDT in this specific use context.
The iPDT has potential to try out scenarios and to get insight in the effects of alternative courses of actions. End-users indicate that they like to use iPDT for training “what-if” scenarios and to be aware of short and longer term consequences of alternative courses of action, to compare alternative decisions and to learn from feedback about different effects (e.g. type, amount and duration of loss; lives and health, economic costs, etc.).
The iPDT used in combination with templates to prepare an exercise and evaluators to give feedback after the exercise, meets the requirements for a training simulation tool.
WP8 – led by CEIS “End-User network”
WP8 mainly included a series of five iterative workshops to present the progress of the work and to integrate crisis management practitioners’ input and expertise into the project's developments. These workshops aimed to allow the target audience of the PREDICT solutions - the crisis management community - to get involved in and to follow the research process. The objective was also for the end-users and the project partners to exchange best practices with their European colleagues on the topics of the project.
WP8 has been carried out in close collaboration with the other research work packages. The results of the workshops have led to the production of 5 deliverables that have been disseminated to the partners with the objective to fed and orient the work of the other WPs.
WP8 therefore included S&T activities (content of the workshop, presentation of project results, analysis of the inputs, transfer of the input to the technical WPs) and supporting activities (practical organisation of the workshop, network management).
The S&T knowledge gained in WP8 falls into three following categories:
• Awareness of European crisis management organisations regarding cascading effects: there are very different levels of awareness and knowledge about domino effects among EU countries. In this regard, the PREDICT project and workshops provided the participants with the opportunity to share common definitions and taxonomy. The concept of cascading effects needs a common definition, a standardised approach of incidents, or levels of alerts, while European countries present very different levels of awareness and expertise on this topic. A common taxonomy would therefore be instrumental to support efficient cross-border collaboration in cascading effect crisis management. In this perspective, PREDICT supports the CIPRNet initiative called CIPedia (www.cipedia.org). Dealing with cascading effects calls for a more holistic approach to crisis management at a European level, by:
o Developing a broader and shared view on risks; initiatives in this field should be promoted such as the work of the European Commission to support the national risk assessment process, methodologies and regional cooperation in the EU.
o Sharing this risk assessment widely to all different stakeholders that can be involved or play a role in crisis response.
• Current practices regarding cascading effect management: it appears that most of the organisations represented during the workshop do not use predictive solutions to handle cascading effects. While cascading effects are increasingly integrated in crisis management plans and procedures, this remains a rather static answer to a dynamic challenge. Crisis management organisations tend to deal with cascading effects more in a reactive way than a predictive one. Information is coming from check-lists and matrices facilitating the identification of dependencies among actors and sectors, and web-based platforms enables information sharing and access to national and sector-specific plans. These tools tend to deal with cascading effects more in a reactive way (by drawing lessons after unpredicted events) than in a predictive one. Thus, plans are rather seen by the workshops’ participants as a static response to a dynamic challenge.
• Use of decision support systems and foresight and prediction tools and assessment of the PREDICT tools: the PREDICT project provide a useful theoretical framework to characterize the evolution of an incident and to identify the dependencies between different sectors and infrastructures. The lack of existing methods to follow the course of an incident, pushes crisis management organisations to rely strongly on lessons learnt processes and not on predictive scenarios. WP8 has highlighted:
o The need and the interest of crisis management professionals for innovative methods and tools to include the management of cascading effects in their work process and procedures.
o The limited added value and the unsuitability of decision support systems such as the iPDT in the response phase. In this phase, crisis management practitioners are focused more on immediate actions and do not have sufficient time to use tools such as iPDT. The risk is to offer too many information and increase the cognitive load of these professionals, to the point where they cannot properly carry out their response actions.
o The importance of focusing on the strategic level in crisis management organisations when developing solutions to manage cascading effects. The strategic level has indeed the necessary distance compared to the field actors and has the role of looking beyond the immediate response to the crisis, through risk analysis and forecasting.
o The value of predictive decision support systems to build and compare scenarios of possible futures addressing a large set of parameters, conditions and data. Such systems would then deliver to strategic decision-makers clear and concise operating instructions that would support the predictive mitigation of the consequences of cascading effects.
Potential Impact:
To date, the PREDICT tools have been brought up to a first experimentation level. Overall, it can be stated that the end-users appreciate the iPDT and its tools especially for training and planning purposes, while limited use is seen to support preparation of an emerging crisis or during a crisis. However, further development is required to make them fit for use for training and planning purposes.
PREDICT tangible results
The PREDICT has resulted in tangible and exploitable results. These results will be used beyond the project for doing business, conducting research and strengthening the positioning of the partners in their field of activity, and in particular in the security sector. The exploitable results of the PREDICT project fall into the following categories:
■ Software tools generated by the consortium developers and intended for further use on their own or in combination with other (iPDT).
■ Research documents describing concepts, specifications for tools, and methodologies which are key channels to support the dissemination of the knowledge and expertise gained in PREDICT.
■ End users’ network and expertise thanks to the multiple engagement activities conducted with practitioners in the field of crisis management, which have been used to ensure the operational applicability of the project’s results and their sustainability beyond the funding period.
Understanding awareness
The PREDICT sessions brought together various categories of end users and from different organisations. Operators do have a very good understanding of the processes within their own organisation and know how to act in case of internal failures. The sessions served to increase awareness amongst them on mutual interdependencies, to understand the consequences of failures for other parties, and what implications their actions for other parties could have.
Roles and functions in crisis management teams
The roles and functions in CM teams were analysed during the system design phase for the three case studies which are in three different countries (Germany, The Netherlands and Finland) (D4.1 [11]). It is very different from one country to another. The roles could be summarized as 1) Decision Maker, 2) Coordinator responsible of the situation awareness 3) Coordinator responsible of the on field operations and 4) the Representatives of the of critical infrastructure operators. From our analysis is can be said that the iPDT is more suited for members of the coordinator team responsible for situational awareness and the coordinator responsible for the on-site operations. The decision maker can be trained with such a tool, but at the day of a crisis, he will receive recommendations from the coordinator responsible for the situational awareness.
iPDT for training of crisis management teams
Scenario sessions proved a good instrument to improve stakeholders’ interactions for a more integrated approach in case of emergencies. The PREDICT tooling has been seen complementary to already existing information systems for improved situational awareness among stakeholders and for the assessment of pros and cons of optional [emergency] actions. Due to integration requirements with information systems that CI-Operators and CM organisation already deploy, and given the time criticality in CM-operations, the IPDT in its current state seems most useful for training environments. For that matter, the PREDICT training methodology provides guidance for (more) effective training programmes. However, specific training functionality need to be added to the existing iPDT.
iPDT integrated in operational environment
In the long run, the integration of PREDICT tooling with other information and decision support tools already deployed, may set standards for data interoperability and data exchange.
Main dissemination activities
During the PREDICT project, the various dissemination activities involved all efforts by which project-related knowledge has been provided to relevant stakeholders in Europe. All partners contributed to their realisation and success, whether through attendance at conferences and expert meetings, contributions to scientific and popular publications etc.
Publications
Communicating on PREDICT findings in the scientific community with papers published in journal during the project had three main impacts:
▪ Contributes to the overall dissemination efforts of PREDICT;
▪ Ensuring that the findings of the project will be useful for future research and projects;
▪ Validating these findings by peers.
In PREDICT, publications have been deliberately aimed at different journals to access both the scientific community and the practitioners community.
The table below presents some of the publications produced during PREDICT.
Mapping of areas presenting specific risks to firefighters due to buried technical networks Author: Amélie Grangeat (CEA)
Paper submitted to a Special Issue of the International Journal of Information Systems for Crisis Response and Management (IJISCRAM) and submitted to ISCRAM Conference 2016
Human vulnerability mapping facing vital services disruptions for crisis managers Authors: Amélie Grangeat (CEA), Julie Sina, Marianthi Theocharidou, Vittorio Rosato, Aurélia Bony
Paper submitted to the ISCRAM 2016 conference.
Submission of an article on the progress of the project for the December 2016 issue Authors: Olivia Cahuzac and Martin de Maupeou (CEIS)
Presentation of the project, its progress and the next steps. Emphasis on the involvement of crisis management practitioners throughout the project
Addressing Cascading Crises in Europe in Crisis Response Journal 10:2 Authors : Olivia Cahuzac (CEIS)
Presentation of the project, its objectives, with a focus on its added value for crisis management practitioners.
Incident evolution framework for crisis situations – Methodologies for the operation time model Authors: Kling, T. and Hakkarainen T. (VTT)
Abstract and presentation related to WP3, in the Nordic Fire & Safety Days 2016. Aalborg University, Copenhagen, Denmark, 16-17 June 2016.
A method for visualisation of uncertainty and robustness in complex long-term decisions. Authors: Hanski, J. & Rosqvist, T. (VTT)
An abstract submitted to the 2016 European Safety and Reliability Conference (ESREL 2016), Glasgow 25-29 September 2016
Attractor-Directed Particle Filtering with potential Fields Author: Patrick De Oude (Thales NL)
Paper presented at the Fusion 2016 conference (www.fusion2016.org) in Heidelberg, Germany.
Elicitation of a utility from uncertainty equivalent without standard gambles Authors: C. Labreuche, S. Destercke, B. Mayag. (Thales FR)
Paper presented during the 13th European Conference on Symbolic and Quantitative Approaches to Reasoning with Uncertainty 2015 (ECSQARU'15). Compiègne, France.
Dependencies and Chain Effects in National Security & Crisis Management Journal Publications from PREDICT partners Marieke Klaver, Eric Luiijf (TNO), Rob Peters, Nico van Os (VRZHZ) and René Willems (TNO) appear in the "Magazine Nationale Veiligheid en Crisisbeheersing" 2015
Conferences and events
The “Cascading Effects” Joint final conference - Brussels (Belgium), 16 & 17 March 2017
PREDICT’s final dissemination activity is the joint final conference where the key project results will be presented at the Cascading Effects Conference,
The conference also presents and discusses results, models and tools from PREDICT, CascEff, CIPRNet, FORTRESS and SnowBall – all projects funded by the European Union’s Seventh Framework Programme.
DOMINO II Conference – Tiel (The Netherlands), 21 & 22 September 2016
On 21 and 22 September 2016, Thales (NL and Fr), ITTI, TNO, and VRZHZ participated in the DOMINO II Conference also held in Tiel and organised by the Dutch First Responder platform, IFV (http://www.ifv.nl).
Several European project took part in this event which features live demos of information sharing systems and presentations on the issues of interoperability in crisis management.
Over 20 crisis management practitioners from all over the Europe, including operators, trainers and high-level decision makers, had an opportunity to get familiar with crucial iPDT functionalities during 3 hands-on sessions.
DOMINO Research and Development – Zwijndrech (The Netherlands), 21 & 22 May 2015
The Dutch First Responder platform IFV organised a joint research and development event about crisis information management during incidents with cascading effects attended by some of the PREIDCT partners. The event - named ‘DOMINO’ - took place May 20, 21st and 22nd 2015.
Over the last five years, the Dutch platform of government emergency professionals – closely related to the National security council and the joint Fire Brigade Commanders of the Netherlands - has been working on issues of interoperability, semantics, geo-information, interpretation and multi-actor data preparation in order to improve the task they have all in common; to improve the information position of the incident managers during a large crisis. Six European FP7 research projects currently addressing this problem including the FP7 PREDICT project took part in this event.
Research Executive Agency (REA) workshop on synergies between FP7 projects – Brussels (Belgium), 25 June 2014
On the 25 June 2014, representatives from the PREDICT consortium (CEIS) took part in a workshop organised by the Research Executive Agency (REA) to explore synergies between FP7 projects related to critical infrastructure protection against natural hazards and cascading effects in crisis situations.
The aim of this meeting was to establish an effective collaboration between different projects, which are dealing with similar issues (critical protection against natural hazards, extreme events LP-HI, and cascading effects in crisis situations), to avoid redundancies and potential duplication of efforts, and to improve the quality of the expected results and boost their impact.
List of Websites:
Website: www.predict-project.eu CEIS
Coordinator:
Dominique Sérafin CEA (Commissariat à l’énergie atomique et aux energies alternatives)
Gramat Center
BP 80200
46500 Gramat
France
dominique.serafin@cea.fr
+33 5 65 10 54 46
Context
“A small cause may have a big effect!” – Decision-makers in crisis situations and Critical Infrastructure (CI) organisations must consider many factors: political, social, legal, cultural, ethical and economic parameters are always to be considered during threat assessments and counter-measures. The increasing complexity of crisis and disasters have mobilized crisis management organisations and motivated the industry and the scientific community to create tools that will be able to help crisis managers to act more quickly and effectively.
The challenge
The aim of the PREDICT project is to provide an innovative set of tools to deal with cascading effects in multi-sectorial crisis situations covering aspects of critical infrastructures. The PREDICT solution consists of a suite of decision support tools (DST) integrating different services facilitating foresight, prediction, communication and eventually decision making in crisis situation. This suite of tools will be based on improved and innovative methodologies, models and software tools.
Project Objectives:
The main objective is to reduce the impact of crisis with cascading effects by:
■ Increasing awareness and understanding of cascading effects.
■ Enhancing the preparedness of crisis management practitioners for cascading effects.
■ Improving their capabilities to respond to crisis.
Main Results:
The most visible results of the project is:
■ A new advanced decision support and training prototype: the iPDT.
But more outcomes have been obtained from the PREDICT project:
■ 3 test-cases across Europe.
■ A cascading effect assessment methodology.
■ A training methodology.
■ Dependencies & interdependencies information.
■ A solid pan-European network of crisis management practitioners.
Project Context and Objectives:
Context
The aim of the PREDICT project was to provide a comprehensive solution for dealing with cascading effects in multi-sectoral crisis situations covering aspects of critical infrastructures.
The PREDICT integrated Tool (iPDT) consists of a suite of decision support tools (DST) integrating different services facilitating foresight, prediction, communication and ultimately decision making in crisis.
This integrated tool is based on improved and innovative methodologies, models and software tools. It aims to increase the awareness and understanding of cascading effects by crisis response organisations, enhance their preparedness and improve their capability to act in case of cascading failures.
Definitions
To use the same terminology and have the same understanding, the PREDICT project selected the following definitions:
Cascading failures: “A cascading failure occurs when a disruption in one infrastructure causes the failure of a component in a second infrastructure, which subsequently causes a disruption in the second infrastructure.” (Rinaldi, 2001)
Critical Infrastructures: “Critical infrastructure’ means an asset, system or part thereof located in Member States which is essential for the maintenance of vital societal functions, health, safety, security, economic or social well-being of people, and the disruption or destruction of which would have a significant impact in a Member State because of the failure to maintain those functions.” (European Council Directive 2008/114/EC)
Objectives
The key objectives of PREDICT were to:
■ Gather and analyse available domain knowledge;
■ Develop a common framework;
■ Create conceptual and executable models of cascading effects and interdependencies;
■ Develop a suite of software tools;
■ Validate the solution through running simulations in the framework of three test cases;
■ Disseminate project results and build appropriate liaisons among stakeholders.
PREDICT project phases
Figure 1 illustrates the different phases of the project to reach our objectives. The first phase of the project consisted in gathering information and performing different analysis to feed the tools’ and models’ developers. This occurred mostly in WP2 & WP3 and in addition, first information from external end-users was gathered in WP8. The second and longest phase of the project was dedicated to the development of the iPDT architecture and of the different models and functionalities. In parallel, the PREDICT test-case were organised by WP7 with a first round of assemblies and two other end-user workshops were also organised with external end-users (WP8). The last phase of the project is the consolidation phase. The iPDT prototype was integrated step by step by clusters. Methodologies were prepared so that the prototype could be tested and discussed during the three PREDICT final test-cases. During the last two end-user workshops, more recent version of iPDT could be presented and tested.
Key results
The 3 main results of the project are as follows:
1. The PREDICT incident evolution framework, which provides a threat quantification methodology allowing for the assessment of cascading effects and the modelling of interdependencies between CIs.
2. Improved versions of tools towards the same objective:
▪ SBR (scenario reasoning): a tool supporting the generation and analysis of most probable set of scenarios;
▪ PROCeed (scenario player): which enables crisis managers to model and simulate crisis scenarios;
▪ Myriad (impacts evaluator): a tool supporting decision by risk-based assessment of the current and predicted situation.
3. The Integrated PREDICT Tool Suite (iPDT) (see Figure 2), which successfully combined these tools to provide a solution enabling decision-makers to generate, run and analyse alternative scenarios of a given crisis, identify crucial dependencies between CIs and act with better understanding of the future.
Furthermore, PREDICT project has delivered a unique set of models and modelling approaches (including objects’ classes, attributes, characteristics and dependencies), which proved to be applicable in three different scenarios (Flooding, Railway Incident and Maritime Incident).
Throughout the PREDICT project, crisis management practitioners have thus followed the progress of work, exchanged best practices with their peers and assisted in defining common needs pertaining to the management of cascading effects. This method of “co-development” ensured that the solutions developed in the project respond to the requirements and constraints of these professionals.
In addition to this, several dissemination activities have been conducted and papers have been published over the course of the project, ensuring that the results of PREDICT are usable and exploitable by the research community.
Project Results:
Project organisation
The PREDICT project was organised in such a way that end-users were deeply involved. They are part of the project at three levels:
■ Three end-user organisations are partner of the PREDICT consortium;
■ End-users member of PREDICT’s Advisory Board ; and
■ Many representatives from end-user organisations across Europe.
Figure 3 illustrates the continuous involvement of end-users in the project activities.
Returning to the work package way of distributing work and responsibilities: in WP7, 3 use cases were developed under the lead of our field partners (UIC – International Union of Railways; VRZHZ – Safety Region South Holland South (NL); and SYKE – Finnish Environment Institute). During the project’s lifetime each year and for each individual use case (scenario) workshops with respective (local) stakeholders have been organized.
■ The workshops in 2014 were beneficial to orient the research activities and to tune the requirements and specifications for the integrated PREDICT tool. The research activities were mostly covered in WP2 (Domain analysis and Requirements) and WP3 (Incident Evolution Framework). The tool development activities started progressively for the backbone of the system that eventually had to integrate the functionalities (WP4) and later for the foresight and prediction functionalities (WP5) and for the decision support functionalities (WP6).
■ The workshops in 2015 and 2016 were held to test and validate advanced mock-ups of individual tools (2015) and the first version of the iPDT integrating the individual tools (2016). The integration of the functionalities was done in clusters and the tool developers worked closely with one of the use case leaders. The functionalities were then not equally developed for the three use cases, but an overall assessment could be done.
Secondly, an end-user network was built in WP8. Five workshops were organized to trigger interest for PREDICT’s objectives and developments and to get further recommendations for iPDT functionalities and applications. The end-user network consisted of a broad set of public crisis managers and infrastructure operators from across Europe. The two final end-user workshops within WP8 allowed the PREDICT partners to discuss the results of the iPDT with different types of end-users and to collect further recommendations for a follow-up of the activities. Next sections elaborate on the setup and context of both approaches, the use cases with end-users, and the end-user network development activities.
WP2 – led by CEA "Domain analysis and Requirements"
Under the title “Domain analysis and Requirements”, WP2 was the baseline activity for the understanding of “Cascading Effects & Resilience” in complex systems through the realisation of a comprehensive assessment. It is a primary step towards the development of PREDICT crisis management methodology and its corresponding integrated DSS. The integrated PREDICT Tool is an end-user oriented tool and it is called “iPDT”.
The main objective of WP2 is to provide a common understanding for the consortium in order to develop a methodology to assess the “cascading effect & resilience” of complex systems. The methodology should admit a given level of interconnectivity and well-defined thresholds. The definition of cascading patterns and the selection of the sets of critical infrastructures will be guided by real crises occurred in the past.
The WP2 was structured in 3 tasks each has its own objective:
■ T2.1; to assess the “state-of-the-art of the R&D in cascade effect & resilience and global modelling”,
■ T2.2; to assess the “security metrics for threats, for systems’ resilience MS&A activities”, and
■ T2.3 to assess the “DSS and predictive tools feature specifications”.
The outputs of WP2 (all three tasks included) are underlying inputs to all other PREDICT’s WPS.
T2.1: State of the art of the R&D activities in Cascade Effect & Resilience and global modelling
The main objective of the task T2.1 is to assess the state-of-the-art on the R&D activities in Cascade Effect & Resilience in crisis situation. The assessment laid down the foundation for a common understanding of the existing views and models in cascade effects and resilience in complex systems. Besides, the assessment has allowed us to determine the complexity of the subject and to evaluate the gap separating between the models and the real end-users’ needs.
The task T2.1 starting point was an extensive survey of the existing approaches of Modelling, Simulation & Analysis (MS&A) of the cascading failures and the CI’s resilience. The survey considered: academic papers; experts reports and letters, books, industrial technical reports, crisis analysis reports and safety/security authorities documents. Out of some 250 different files covering about 10 recent years, at the end 122 have been considered in the assessment because of their relevance. We recall that out target in methodologies, models and tools in cascade effects & resilience of complex systems. A distinction between qualitative and quantitative approaches has been used as an a priori filter.
The state of the art was categorised in main areas that we deemed relevant for PREDICT tasks: threats identification and classification, dependencies and interdependencies between critical infrastructures, crisis management and decision support system (DSS) specifications. Some sub-categories have equally been underlined: failure data and databases identification, failures propagation, threats detecting and periodic monitoring, cascade dynamic modelling and resilience metrics. In addition, D2.1 put some light on the state of the art in taxonomy corresponding to threat, critical infrastructure, and crisis codification.
The assessment covered three macros areas. The 1st macro area is related to qualitative activities which will allow us identifying and classifying the most plausible threats considering past crisis situations. The 2nd macro area is related to the identifying, assessing and commenting the existing quantitative methodologies and models in cascading assessment. The 3rd macro area is dedicated to the development of a taxonomy for threats based on collected past and the running R&D effort and the harmonization of existing taxonomies relative to the global security. The assessment considered the crisis management end-users practices and requirements, as well.
T2.2: Security Metrics for threat and system’s resilience MS&A
Generally, threat identification and characterisation is a first act in almost all MS&A of cascade effects. A pertinent identification and characterisation of threats would necessarily be based on the use of the most appropriate security metrics. The starting point for the development of appropriate security metrics was a comprehensive state-of-art analysis of different existing metrics in different disciplines. This allows a better determination the classes of causes, consequences and sequences.
Task T2.2 aims at addressing those criteria of appropriate security metrics from threat identification & characterisation standpoint. The demarche was based on recalling the current learnt-lessons in some representative fields where some systems (and systems of systems) had been exposed to severe threats (sometimes catastrophic) and the threats had to be managed. Some learnt-lessons could successfully be translated in codes, norms and regulations and subsequently received worldwide recognitions. Others are still in a pre-maturation phase with limited local recognition. Few others are kept at the stage of local good practices with no external visibility.
Equal efforts were put is the task to cover the two main categories of threats: Natural and man-made ones. Only threats with adaptability-features have been excluded: terrorist actions, sabotages and wars.
The deliverable addressed the generic criteria and specifications of the security metrics that could be considered as appropriate for threats identification & specification.
T2.3 DSS and predictive tools feature specifications
Task 2.3 is the “tools feature specifications”. The final specifications of the iPDT will be set in the technical work-packages WP4-5-6, sticking as much as possible to the proposed features recommended by T2.3. The feedback development experience from the other WP’s will be considered to update these recommendations-specifications if necessary. The updating is a dynamic process in order to maintain and to guarantee a common understanding of the cascade effect MS&A, throughout the whole life of the project. The specifications have been critically assessed by WP4 “System Design & Iterative Integration”, as WP4 is to develop and implement the architecture of the iPDT.
WP2 has equally associated end-users in the critical assessment of its recommendations & specifications through a workshop on “MS&A of cascade effects-WP2 findings” (WP8). The end-user expressions & needs for the iPDT have been collected within the consortium and processed. During this workshop, various end-users in all CIP fields provided their inputs. The results of the workshop has been analysed by the project’s tool provider and the system’s architect (THALES, TRT-NL, iTTi, VTT and Fraunhofer). Furthermore, the scenario methodology task, which overlaps with T2.2 has been completed within T2.3.
Task 2.3 was structured in 4 complementary topics.
The first gave a general overview on “Decision Support Systems and Predictive tools”, based on D2.1 assessment outputs. This put the focus on the major feature specifications.
In the second topic, T2.3 assessed the end-users requirements and specifications, from deliverable D8.1 [PREDICTD81] and subsequently updated its recommendations & specifications.
In the third topic, WP2-T2.3 focused on the predictive capability of iDPT.
The last topic is dedicated to the feature specification of “CIPRNet methodology to create crisis scenarios”. Within this topic, examples of scenarios in different crisis situations were treated. These crisis situations covered: flooding, train transporting hazardous substance derailment, Nordic winter storm, coast submersion, earthquake and large forest fire.
Outputs to other PREDICT-WPs
PREDICT-WP2 provides a baseline for understanding cascade effects and resilience in Critical Infrastructures facing given threats. This comprehensive assessment of cascading effects and resilience of systems is a primary step towards the development of the appropriate PREDICT cascade effect Modelling, Simulation and Analysis (MS&A) PREDICT predictive tool suite “iPDT”. WP2 came up with a set of conclusions and recommendations to guide the design of PREDICT Tool Suite for cascade effect MS&A. WP2’s major outputs to other PREDICT WPs were listed in a table in the deliverables and project periodic report.
WP3 – led by VTT "Incident evolution framework"
The main S&T results/foregrounds of WP3 “Incident evolution framework” are the seven-step methodology for assessing cascading effects, the cascade probability function providing a methodology to assess the likelihood of cascading effects, the modelling approach for threat quantification revealing cascading effects of interdependent critical infrastructures (CIs), and further development of the Stochastic Operation Time Model (SOTM) estimating the uncertainty related to human factors.
Seven-step methodology was developed for assessing cascading effects. It describes the steps that a crisis management organisation should perform in order to identify the cascading effects for a specific threat scenario. The steps include 1) identifying the threats to be considered, 2) identifying the CIs in the region, 3) identifying the key CI elements, 4) characterising the vulnerability of the key CI elements to the threat, 5) assessing the first order impact of the threat on the CI elements, 6) describing the dependencies between the CI elements in the region (i.e. describing the required input and output of all key CI elements, distinguishing between the different modes of operation, and including the temporal and spatial factors), and 7) assessing the CI cascading effects. The seven-step methodology can be used by end-user experts when figuring out crisis situation scenarios and providing input data and estimates.
Cascade probability function is a mathematical formulation of the seven-step methodology, showing how the steps can be performed using time-dependent, stochastic input. The modelling framework for the cascading effects of interdependent CIs integrates a threat function which reflects the time evolution of the threat scenario, a vulnerability function which describes the vulnerability of the CI elements for a specific threat intensity, and a dependency function which describes the dependency between the different CI elements in the specific area of consideration. This procedure results in a cascade probability function, providing a methodology to assess the likelihood of cascading effects.
The seven-step methodology is used for providing input data and estimates, and the cascade probability function then generates the final estimates for decision makers. Thus, the seven-step methodology can be seen as “a user interface” and the cascade probability function as “an engine” for calculations.
Modelling approach for threat quantification was developed using the PREDICT case studies as development environment. In the modelling approach, an accident scenario map, locating the initiating events and CIs, is defined and a hexagonal grid is laid on the map. The relevant CIs for each hex are identified, and their interdependencies and vulnerabilities are defined in the model. A reference point is chosen for each hex. The threat function results (provided e.g. by a separate modelling tool) on the reference point are then applied to all CIs in the hex. The initial failure times (i.e. not considering the interdependencies) of the CIs are determined. The final failure times are determined taking into account the interdependencies between the CIs, and the cascading effects are assessed. Thus, the modelling reveals the cascading failures in which a CI is lost due to a failure of another CI.
The modelling approach for threat quantification requires that sufficient data on relevant CIs and their locations, interdependencies, vulnerabilities and initial failure times due to threat functions is available. Depending on the input data available, model parameters can be implemented either as single-value point estimates or as probabilistic distributions. The results can be presented on a timeline to illustrate the incident evolution.
Threat quantification modelling can be utilised in crisis management in both the preparedness and training phase and the response phase. In the preparedness and training phase, the modelling illustrates the progress, influencing factors and potential cascading effects of accident scenarios. It gives guidance for the planning of emergency response by revealing the CIs that are important to protect in order to mitigate or prevent the escalation of the accident. To support contingency planning, the threat quantification modelling can reveal crucial vulnerabilities and dependencies which should be eliminated or mitigated to strengthen the resilience of the CIs. In the response phase, scenarios pre-examined in the preparedness phase can be used as references to support decision making. New simulation results on the threatening phenomena can be input to the threat quantification model to correspond the real-life accident, and predefined vulnerabilities and dependencies can be adjusted by expert judgement using the information obtained from the evolution of the crisis.
Stochastic Operation Time Model (SOTM) was originally developed under the Finnish national research programme for nuclear power plant safety. This model has been further developed in PREDICT WP3 to take into account the uncertainty associated to human actions, and the consequences of delays which cause failure risks in response operations. The novelty of the methodology is related to the combination of temporal and reliability aspects: failures in human operations are not considered as final stages but rather as additional time delays of a specific action. Furthermore, the development work aimed at making SOTM feasible for and implementable to iPDT, the integrated PREDICT decision support tool.
SOTM is based on the assumption that human operations can be described as time delays and possible additional delays due to unexpected factors. In WP3, a mathematical description of the model was presented and a generalised methodology for the construction of the model was developed. The methodology consists of modelling steps for the stochastic estimation of operation times, and timeline charts to visualise the action and communication processes of response procedures. Carrying out Monte Carlo analysis, a probability distribution for the total duration of response operations is obtained. By comparison to the temporal development of the crisis, it can be assessed with which probability the response actions can prevent the escalation of the situation. The modelling gives insight into interdependencies of various phenomena and activities of different actors, and can reveal bottlenecks and contradictions in processes and organisations.
A set of methodologies for the specification of model parameters related to human operations was studied, and the use of the concept of performance-shaping factors (PSFs) was suggested. Based on the applicability and availability of the modelling options in the iPDT development suite, agent modelling was chosen for further investigation and for implementation.
The seven-step methodology, the cascade probability function (including threat, vulnerability and dependency functions), the threat quantification methodology with its various phases, and the stochastic operation time model with the set of methodologies for the specification of model structures and the parameters related to human operations have been forwarded to WP5 and WP6 for implementation into iPDT.
WP4 – led by Fraunhofer “System Design & Iterative Integration"
The main objective of WP4 was to develop a software architecture for integrating FPT (developed in WP5) and DST (WP6) into one coherent technical solution for cascading effects analysis in civil domains. Based on this architecture, the iterative system integration process needed to be conducted by WP4 to achieve the final integrated system. These two objectives have been successfully achieved by WP4 (together with other WPs) in PREDICT. The main results/foregrounds are:
• Software architecture for system integration based on the MCRI concept. WP4 proposed an architectural approach for integrating information systems used in a typical crisis management process, which consists of three parts: situational awareness, foresight and prediction, decision making. This architecture tries to combine modern software engineering best practices with the specific requirements in crisis management process. RESTful mock-up services based on modern Service-Oriented Architecture (SOA) are the essential parts of the integration architecture. Reverse- proxy based solution provides a flexible runtime environment for hiding the technical details of different system implementations. Special design consideration is also given to integrate spatial data into the crisis management process. To maximise the system design flexibility, software containers are used to provide flexible wrappers for the real implementation. This architecture is proven to be effective for the integration work in PREDICT.
• The iterative integration approach. Together with the integration architecture, the iterative integration approach is also one of the major outputs of WP4. Working with partners from different organisations on the same software project can be difficult, especially when it comes to integrating new system features and providing system maintenance. It can yield unwanted dependencies and slow down the software development process. An iterative approach of system integration can be separated into three stages:
1. Defining specification and requirement of the service. This includes developing use cases, formal specification, etc.
2. Writing service mock-ups and deploy them to the server for automated testing. After this stage, all unit tests should pass as required in classical Test-Driven Development (TDD).
3. Iteratively replace mock-ups by real implementations. Each time, if a service mock-up is replaced, all unit tests must be executed to guarantee that the service implementation meets the requirements defined in the specification.
The first step required a complete specification description of each RESTful endpoint including:
1. Uniform Resource Locator (URL) - a unique ID of the service.
2. Request method type - indicating the character of the service whether it is for reading, writing or deleting operations.
3. Header information - some meta information that is not suitable to be encoded in the URL.
4. Payload - additional information that is two large to be encoded in the URL.
5. Expected responses - the result delivered by the service implementation.
Finally, after all the mock-ups were replaced by real implementation, it is always a good practice to have some high-level tests running like functional tests and even human-guided tests.
• The integrated system iPDT for cascading effects analysis. iPDT is the major technical output of WP4 with the tools developed in WP5 (PROCeed) and WP6 (MYRIAD and SBR). Regarding the result of the integration, iPDT is characterised as a system that delivers two ranges of functionality. First, it delivers system-level user functions for accessing iPDT and its component systems (like a user login) and a developer access for monitoring and maintenance. Second, the main added-value of iPDT is that it enables its component systems to use the other components as additional services to provide integrated functionality. There is no specific iPDT GUI, but, for instance, in “iPDT mode” the PROCeed GUI provides more functions than in “stand-alone mode” and can display information generated by the other system components MYRIAD and SBR. iPDT is implemented in a distributed fashion. The physical core of iPDT is a reverse proxy server hosted at Fraunhofer. It connects to the partner’s servers that run the other system components (FPT, DST, middleware, external tools). The integrated system has been demonstrated at three end-user workshops organised by WP7. Feedbacks coming from the end-users during these workshops provide invaluable information for the future improvement of iPDT.
WP5 – led by iTTi "WP5 “Foresight and Prediction Tools”
There were two main objectives of the WP5:
1. Design and develop of the Foresight and Prediction component for the iPDT solution ; and
2. Design and develop appropriate interfaces for the WP5 and WP6 tools communication
Both objectives has been fully achieved. The most important technical results are:
a. Simulation Engine Development
The concept of the PREDICT – FPT simulation was to follow a stateless approach, which is partially based on the Discrete – Event Simulation Concept. Thus there is no continuous simulation being processed, as there is a simulation on demand (step by step approach) implemented. Hence the interactions (also referred as ‘actions’) with the user and external tools are required. Following interactions are distinguished in the PREDICT – FPT simulation: (i) sending and receiving simulation data from/to external tool, (ii) ad-hoc interactions with the user or user’s action, and (iii) execution of a simulation module of the PREDICT – FPT.
The simulation in the PREDICT project is a process of interpreting and processing chronologically ordered events and calculating the future course of events provided by the simulation engine and supported by other PREDICT functionalities (e.g. SBR) and simulation modules (e.g. rule engine).
b. Rules Engine Development
In order to effectively simulate the evolution of the crisis, a Rule Engine has been developed as part of the Simulation Engine. The rules on the conceptual side are based on Drools (www.drools.org - a business rule management system) which is a well-known solution dedicated for a business process management. The syntax of the rules is based on MVEL (MVFLEX Expression Language) – a hybrid dynamic/statically typed, embeddable Expression Language and runtime for the Java Platform. The rules on the conceptual side are based on Drools business rule management system, which is a well-known solution dedicated for a business process management. The syntax of the rules is based on mvel2 (MVFLEX Expression Language) – a hybrid dynamic/statically typed, embeddable Expression Language and runtime for the Java Platform.
c. Development of PREDICT-FPT Modules
The PREDICT-FPT component follows the concept of object-oriented modelling and object-oriented design. Consequently, data storing follows a concept of object-relational mapping. Development of the PREDICT-FPT component was based on the following modules: (i) Users' Actions Module, (ii) Object Management Modules, (iii) Rules and Interactions Module and (iv) Scenario Management Modules. Each module is controlled by the Simulation Engine which is interfaced with the PROCeed tool and generates simulation results. The generated results are available for the iPDT, which is the main data broker available for all the modules and tools running under PREDICT–FPT and PREDICT–DST subsystems.
d. Development of Interfaces for Decision Support Tools
The PREDICT-FPT component is responsible for managing the simulation (scenario evaluation) process. The communication between different components is based on RESTful APIs of the tools. The communication between the tools is synchronous and the initiator of the communication is the PREDICT-FPT component. Exceptionally other data sources treated as sensors or user interaction are based on the asynchronous communication. It is of crucial importance that a single event describes the change of an object and each message (describing an event) consists of the object identifier, time of the event occurrence and event type (i.e. predefined; created by user; generated by external tool; detected by sensor). The stateless simulation approach is used within the data integration.
e. Technical implementation of the WP3 Incident Evolution Framework
WP3 provided a detailed description of the 7-step methodology for the cascading effects identification and assessment. Several recommendations for the technical implementation were provided as well. PREDICT-FPT followed the methodology itself and provided recommendations. Thus crucial elements of the cascading effects identification process were fully implemented in the software. Thus the user is enabled to conduct set of actions within the tool, which lead to cascading effects identification and assessment. Following activities are considered to be crucial from the cascading effects identification perspective:
(i) Identification of the threats to be considered at the particular location;
(ii) Identification of the Critical Infrastructure located in the region;
(iii) Identification of the key CI elements;
(iv) Implementation of the vulnerability function and first order impact assessment;
The interconnections between the 5 modules developed is illustrated in Figure 5.
WP6 – led by THALES “Decision Support Tools”
The objective of this WP was to predict the range of possible consequences of each decision, and to determine, as accurately as possible, the probability that each credible consequence will occur.
WP6 has fully achieved his goals. The main results are:
• Establish formal and mathematical models of risk indicators
• Establish a hierarchical task analysis (HTA) describing the goals and tasks involved in the decision making processes
• Multi Criteria Decision Making approach, based on a methodology that models and solves the MCDM problems with imperfect information
• A tool proposing a list of actions mitigating the risks to achieve the goals of the expert
• A Dynamic Expert Integration Network that supports a systematic combination of fusion and analysis services into complex assessment systems with humans in the loop
• A tool that facilitate description of uncertainties by developing alternative scenarios
• These achievements are documented in deliverables and demonstrated through the tools in the iPDT suite.
The results of each task is summarized in the following paragraphs:
Task 6.1 Risk Indicators Requirement and Specifications
The deliverable D6.1 summarizing the work done during this task. It states the “Requirements and specifications report of Decision Support tools” provides a structured framework to construct an aggregated risk, fit for use in cascading effect situations. In a more detailed way, this deliverable presents, first, a method: the Multi-Criteria Decision Making Approach (MCDA), to identify the risks indicators requirements and specifications. Next, this document presents the hierarchy of point of view based on the three components of risk assessment for cascading effects: likelihood, impact and topology of the network and defines theses points of views thanks to the key relevant questions answered by the end-users. The document ends by describing how we can operationalize the points of views with metrics and aggregations adapted to the Predict needs with a mathematical structure allowing formalizing and defining of “non-naturally numerical” impacts, which can then be used to contribute to the risk level, and focused on criticality and geography of the affected areas, as well as impact of time.
Task 6.2 Cognitive analysis to support decision tools specifications
In order to do the cognitive task analysis a simulation of a ROT (regional operational team) situation. A working session was organised during a test-case workshop on October 15th, 2015 in Dordrecht. The participants of the ROT simulation were asked to consider the flooding case and the possibility of an evacuation. During this simulation we made an initial attempt to do a hierarchical task analysis (HTA) describing the goals and tasks involved in the decision making processes. Due to the dynamics of the discussion, it was difficult to obtain the HTA during this event. The simulation was recorded on video for further post analysis. Based on the video material another workshop was organized on November, 26th in Delft to analyse results of video material & transcripts. Additionally, historical accident reports related to railways were analysed. We were able to specify an HTA tree, test it for the three test-cases and identify where the PREDICT tools could be used in the decision making process of ROT members. The details of this research are available in Deliverable D6.2.
Task 6.3 Multi-criteria decision making under uncertainty
This task is the topic of D6.3. It describes the technical aspects of the risk assessment module which is based on multi-criteria decision making under uncertainty. This is the core of the “impacts evaluator” of the iPDT. It provides an analysis to choose a multi-criteria decision making approach by identifying the functional requirements that shape the module, and derives a suitable MCDM model and its functionalities. It also tackles the definition and handling of uncertainty in the module: imprecise values of the input, of the parameters of the decision model (e.g. weights, value functions) itself, partial relevance of the sectors in each use-case. The multi-criteria decision model shall thus be reshaped automatically to focus only on the relevant sectors, using for instance expected utility, pessimistic rule or min-max regret.
Task 6.4 Uncertain decision making tool to support short or mid-term planning
Deliverable D6.4 summarizes the approaches developed to assess the impact and risk of cascading effects, which have been implemented in the MYRIAD tool. This includes conversion of scenario trees into individual sector assessments, and identification of optimal sequence of actions (i.e. decisions). Three methods are proposed: identification of the “most probable” scenario (adapted to PROCeed workflow); computation of the expected risk given uncertainties; computation of the min-max regret (minimization of the regret). The implementation of the MYRIAD tool is also described in D6.4 and shows how the user can modify the preference information to adapt to a new use-case.
Task 6.5 Dynamic Expert Integration Network
The Dynamic Expert Integration Network (DEIN) is an information sharing tool for experts to quickly share critical information during large scale crisis events. This tool which is not fully integrated to iPDT is more operational oriented. DEIN is a special instance of the Dynamic Process Integration Framework (DPIF) that focusses only on integration of human expert instead of automated processes. Since DPIF is a SOA the communication in DEIN is service-based. This means that all communications between different experts is facilitated through the expert’s provided and required services. In that sense, it is very different from other type of communication systems, such as the telephone, e-mail, chat, etc., where you have to know the contact details of the person you want to communicate with. In DEIN communication is established based on the type of information you, as an expert, can provide or need. The person (typically another expert) that needs or can provide critical information does not necessarily have to be known in order to communicate since the communication is service-based. For example, a service can be provided by an organization with multiple experts that can actually provide this service. In this way redundancy is automatically build in and different experts can provide time sensitive information quickly.
Task 6.6 Scenario-based Reasoning Tool
The Scenario-based reasoning (SBR) tool consists of a “frontend” and a “backend” part (see Figure 7). The frontend solely consists of a browser-based GUI that uses RESTful web services to communicate with the backend (see the communication channel depicted as arrows in Figure 7).
With the help of SBR frontend the end-users can view different aspects of CIs (like states, location, etc.) and other type of objects like schools, hospitals etc. that are involved in different scenarios for investigation. For each CI/object in SBR, the user can view the likelihood of CI/object failure at a certain point in time. Additionally, the graphical user interface supports the visualization to up to three alternative developments (most likely, worst-case and best-case scenarios) for a given scenario concerning the different failure times of CI. These possible scenarios are directly presented in the timeline component (see the bottom part of Figure 8 annotated with an orange rectangle) of the GUI. By clicking on the red or green icons shown in the timeline (see Figure 8), the user can easily find the corresponding CI/object on the map view.
Different scenarios can be inspected by setting observation about the CI/objects shown in the map (see the green block in Figure 8). For example the user can tell the system that a certain CI failed at a certain time by setting the value of CI state to not working. Based on this observation, the SBR engine is able to compute the likelihoods of failure for the other CIs/objects by considering the dependency model embedded in the test case. This allows the user to inspect the failure time of different CI/objects under different conditions (what-if exploration).
WP7 – led by TNO “Prototype and Testing Validation”
The purpose of Work Package 7 (WP7) within the PREDICT project is to appreciate (elements of) the iPDT with end-users who are acting in the three PREDICT test-cases. For that matter, a series of engagements with end-users has been organised to identify their requirements, to present first versions of elements of the iPDT, and to get an appreciation of the iPDT in its final stage with recommendations from end-users for the way ahead. The specific use cases were developed under the lead of our field partners (UIC – International Union of Railways; VRZHZ – Safety Region South Holland South (NL); and SYKE – Finnish Environment Institute).
During the project’s lifetime each year and for each individual use case workshops with local stakeholders have been organized.
■ The workshops in 2014 were beneficial to orient the research activities and to tune the requirements and specifications for the integrated PREDICT tool.
■ The workshops in 2015 and 2016 were held to test and validate advanced mock-ups of individual tools (2015) and the first version of the iPDT integrating the individual tools (2016).
The WP7 S&T outcome can be divided into three parts:
■ General approach → for testing the PREDICT results
■ Validation results → through case studies
■ Training methodology → to support end-users in developing a training using PREDICT results
General Approach
A framework for prototype testing in case studies allowing for the evaluation of the applicability and added value by the PREDICT tools and methods suite has been developed. It was decided to have the end-user feedback on the tools in a two-step approach. In the first step concepts of the tools in mock-up versions will be presented to the end-users. With the feedback then received, the tools have been further developed into prototype versions and integrated in the PREDICT tool. This gives the maximum opportunity for end-user to influence the development of the tools. The final workshops then have the purpose of assessing the value added of the iPDT in the crisis management process.
A set of evaluation criteria has been offered based on which assessment of the added value, both from operational and organizational perspective. To assess the iPDT potential the following criteria were taken into account:
Functional: does the tool provide the functions required for managing cascading effects?
Information: does the tool provide the information required for managing cascading effects?
Organizational fit: does the tool fit the organization of end-users?
Overall acceptance: do end-users intend to use this technology?
Evaluation schemes and forms have been prepared for a uniform evaluations by individual participants to workshops and for condensed workshop reporting.
Validation results
Two series of workshops were planned. For the first workshop series the tool developers have, in close alignment with case leaders, prepared a mock-up for their tools for demonstration in the selected scenarios. The conclusions of these workshops can be summarized as follows:
■ The tool(s) have training value for the operators, and are considered useful for planning purposes;
■ For the operational phase (hot phase of the crisis), end-users don’t see a need for additional tools;
■ The possibility to run different crisis situation scenarios was much appreciated by the users;
■ Regular meetings with the case leader and the different stakeholders are essential to revisit the value added of tool and its results (usefulness, reliability);
■ Individual quality and value of tools should have the highest priority; and only then to seek for integration in a suite of tools;
■ It is important for the end-users to know what kind of models are implemented in the tool.
Referring to the above conclusions it has been decided to not use iPDT in an operational setting during the PREDICT project timeframe. The second series of workshops therefore was developed for a training setting and to use the iPDT as supporting tool for this training. During the workshop in Dordrecht a PREDICT team member supported the training of the responsible crisis management team with cascading effect information based on iPDT. At the Paris workshop (UIC), the iPDT has been demonstrated to stakeholders from a far broader community than only crisis response organisations. At the Helsinki workshop a training regarding cascading effects has been planned at which the iPDT has been demonstrated and a hands-on session has been held. All workshops offered the possibility to extensively discuss the usefulness and the necessary add-ons needed for future use in practice.
The overall results of the second workshop series confirmed the earlier outcome that end-users appreciate the iPDT rather for planning and training purposes, with little or no use for preparation in case of an emerging crisis or during a crisis. They could foresee that the IPDT may work in their organization. This would however require adjustment of the procedures for planning and training and optimization of the iPDT for this use.
Users positively appreciate the functions and information in support of creating awareness of failing critical infrastructures and, to a lesser degree, for understanding cascading effects. Since decision support (MYRIAD component) was not yet fully integrated, end-users were not able to appreciate support for making decisions and missed information about the effects of alternative courses of action.
The discussions provided good insights for improving iPDT. The system has potential to try out scenarios of cascading effects and to get insight in the effects of alternative courses of actions. End-users like to use the iPDT for training “what-if” scenarios; to learn about short and long term consequences; to compare alternative courses of action; to learn about the impact on society and whom to inform and coordinate with. To realize this potential the following is recommended:
■ Integrate and assess the modules for supporting decisions about alternative courses of action.
■ Optimize iPDT for training
■ Optimize iPDT for planning (risk analysis and management)
■ Identify stakeholders and end-users to test new functions for the planning and training use case
■ Develop guidance for using iPDT for training and planning purposes
■ Develop databases with scenarios for training and planning
■ Use recommendations from end-users to develop iPDT for training
Finally, participants at the workshops underlined their interest in hands-on experience with the iPDT in assessments of next versions of the iPDT after the project.
Training methodology
A generic approach for training methodology is based on training requirements, supporting training tools, the actual training and evaluation of training results. This can be applied to various levels of trainees (operators, technicians, trainers and even tool developers).
A stepwise methodology (Event Based Approach to Training - EBAT) have been provided to develop such necessary training and examples showed how this method can be used for training of crisis managers using iPDT. The most important recommendations from the EBAT approach are:
■ To develop learning experiences or learning tasks based on learning goals and not on operational tasks
■ To specify and introduce specific events in any training session that elicit the behaviour that contribute to the acquisition of necessary knowledge and skills
■ To create a meaningful learning environment for trainees where they can experience consequences of correct and incorrect behaviour, and receive feedback on how to improve their behaviour.
The EBAT approach has been used to prepare one of the workshops in the second series. Practical guidelines to develop exercises using a structured approach (including templates) have been proposed and used. This structured approach will support the exercise developer and tool experts to work together to develop an exercise to gain awareness, understand and manage cascading effects supported by the capabilities offered by iPDT. The possibilities of iPDT in combination with this structured approach as a tool for training are evaluated:
■ For iPDT to be used for other scenarios a more extensive validation exercise is necessary for a fully developed iPDT
■ Pre-defined learning objectives can be reached when using the offered structured approach supported by templates
■ The iPDT evokes a learning process: decisions and actions must be made and it is possible to simulate the effect of different decisions in the tool.
■ During exercises supported with iPDT, evaluators should be used and instructed to provide feedback.
■ Actual integration in the iPDT with connectivity to already operational systems leads to consistency over multiple systems
Based on the results, it can be concluded that iPDT is suitable as a tool for training, especially when used in combination with the templates and dedicated evaluators to provide feedback. iPDT in combination with the structured EBAT approach to exercise development, supported by the templates, forms a strong ensemble for training purposes.
The end-users foresee that the iPDT may work in their organization for training or planning. Adoption would be facilitated by a clear business case for the iPDT in which the benefits and costs of using the iPDT are formulated and supported with evidence. Adoption would also be facilitated by procedures for planning and training that are compatible with an improved version of the iPDT.
Optimization for planning requires specification of risk analysis and management in planning processes (e.g. goals, activities, results in existing procedures) and identification of requirements of end-users for using the iPDT in this specific use context.
The iPDT has potential to try out scenarios and to get insight in the effects of alternative courses of actions. End-users indicate that they like to use iPDT for training “what-if” scenarios and to be aware of short and longer term consequences of alternative courses of action, to compare alternative decisions and to learn from feedback about different effects (e.g. type, amount and duration of loss; lives and health, economic costs, etc.).
The iPDT used in combination with templates to prepare an exercise and evaluators to give feedback after the exercise, meets the requirements for a training simulation tool.
WP8 – led by CEIS “End-User network”
WP8 mainly included a series of five iterative workshops to present the progress of the work and to integrate crisis management practitioners’ input and expertise into the project's developments. These workshops aimed to allow the target audience of the PREDICT solutions - the crisis management community - to get involved in and to follow the research process. The objective was also for the end-users and the project partners to exchange best practices with their European colleagues on the topics of the project.
WP8 has been carried out in close collaboration with the other research work packages. The results of the workshops have led to the production of 5 deliverables that have been disseminated to the partners with the objective to fed and orient the work of the other WPs.
WP8 therefore included S&T activities (content of the workshop, presentation of project results, analysis of the inputs, transfer of the input to the technical WPs) and supporting activities (practical organisation of the workshop, network management).
The S&T knowledge gained in WP8 falls into three following categories:
• Awareness of European crisis management organisations regarding cascading effects: there are very different levels of awareness and knowledge about domino effects among EU countries. In this regard, the PREDICT project and workshops provided the participants with the opportunity to share common definitions and taxonomy. The concept of cascading effects needs a common definition, a standardised approach of incidents, or levels of alerts, while European countries present very different levels of awareness and expertise on this topic. A common taxonomy would therefore be instrumental to support efficient cross-border collaboration in cascading effect crisis management. In this perspective, PREDICT supports the CIPRNet initiative called CIPedia (www.cipedia.org). Dealing with cascading effects calls for a more holistic approach to crisis management at a European level, by:
o Developing a broader and shared view on risks; initiatives in this field should be promoted such as the work of the European Commission to support the national risk assessment process, methodologies and regional cooperation in the EU.
o Sharing this risk assessment widely to all different stakeholders that can be involved or play a role in crisis response.
• Current practices regarding cascading effect management: it appears that most of the organisations represented during the workshop do not use predictive solutions to handle cascading effects. While cascading effects are increasingly integrated in crisis management plans and procedures, this remains a rather static answer to a dynamic challenge. Crisis management organisations tend to deal with cascading effects more in a reactive way than a predictive one. Information is coming from check-lists and matrices facilitating the identification of dependencies among actors and sectors, and web-based platforms enables information sharing and access to national and sector-specific plans. These tools tend to deal with cascading effects more in a reactive way (by drawing lessons after unpredicted events) than in a predictive one. Thus, plans are rather seen by the workshops’ participants as a static response to a dynamic challenge.
• Use of decision support systems and foresight and prediction tools and assessment of the PREDICT tools: the PREDICT project provide a useful theoretical framework to characterize the evolution of an incident and to identify the dependencies between different sectors and infrastructures. The lack of existing methods to follow the course of an incident, pushes crisis management organisations to rely strongly on lessons learnt processes and not on predictive scenarios. WP8 has highlighted:
o The need and the interest of crisis management professionals for innovative methods and tools to include the management of cascading effects in their work process and procedures.
o The limited added value and the unsuitability of decision support systems such as the iPDT in the response phase. In this phase, crisis management practitioners are focused more on immediate actions and do not have sufficient time to use tools such as iPDT. The risk is to offer too many information and increase the cognitive load of these professionals, to the point where they cannot properly carry out their response actions.
o The importance of focusing on the strategic level in crisis management organisations when developing solutions to manage cascading effects. The strategic level has indeed the necessary distance compared to the field actors and has the role of looking beyond the immediate response to the crisis, through risk analysis and forecasting.
o The value of predictive decision support systems to build and compare scenarios of possible futures addressing a large set of parameters, conditions and data. Such systems would then deliver to strategic decision-makers clear and concise operating instructions that would support the predictive mitigation of the consequences of cascading effects.
Potential Impact:
To date, the PREDICT tools have been brought up to a first experimentation level. Overall, it can be stated that the end-users appreciate the iPDT and its tools especially for training and planning purposes, while limited use is seen to support preparation of an emerging crisis or during a crisis. However, further development is required to make them fit for use for training and planning purposes.
PREDICT tangible results
The PREDICT has resulted in tangible and exploitable results. These results will be used beyond the project for doing business, conducting research and strengthening the positioning of the partners in their field of activity, and in particular in the security sector. The exploitable results of the PREDICT project fall into the following categories:
■ Software tools generated by the consortium developers and intended for further use on their own or in combination with other (iPDT).
■ Research documents describing concepts, specifications for tools, and methodologies which are key channels to support the dissemination of the knowledge and expertise gained in PREDICT.
■ End users’ network and expertise thanks to the multiple engagement activities conducted with practitioners in the field of crisis management, which have been used to ensure the operational applicability of the project’s results and their sustainability beyond the funding period.
Understanding awareness
The PREDICT sessions brought together various categories of end users and from different organisations. Operators do have a very good understanding of the processes within their own organisation and know how to act in case of internal failures. The sessions served to increase awareness amongst them on mutual interdependencies, to understand the consequences of failures for other parties, and what implications their actions for other parties could have.
Roles and functions in crisis management teams
The roles and functions in CM teams were analysed during the system design phase for the three case studies which are in three different countries (Germany, The Netherlands and Finland) (D4.1 [11]). It is very different from one country to another. The roles could be summarized as 1) Decision Maker, 2) Coordinator responsible of the situation awareness 3) Coordinator responsible of the on field operations and 4) the Representatives of the of critical infrastructure operators. From our analysis is can be said that the iPDT is more suited for members of the coordinator team responsible for situational awareness and the coordinator responsible for the on-site operations. The decision maker can be trained with such a tool, but at the day of a crisis, he will receive recommendations from the coordinator responsible for the situational awareness.
iPDT for training of crisis management teams
Scenario sessions proved a good instrument to improve stakeholders’ interactions for a more integrated approach in case of emergencies. The PREDICT tooling has been seen complementary to already existing information systems for improved situational awareness among stakeholders and for the assessment of pros and cons of optional [emergency] actions. Due to integration requirements with information systems that CI-Operators and CM organisation already deploy, and given the time criticality in CM-operations, the IPDT in its current state seems most useful for training environments. For that matter, the PREDICT training methodology provides guidance for (more) effective training programmes. However, specific training functionality need to be added to the existing iPDT.
iPDT integrated in operational environment
In the long run, the integration of PREDICT tooling with other information and decision support tools already deployed, may set standards for data interoperability and data exchange.
Main dissemination activities
During the PREDICT project, the various dissemination activities involved all efforts by which project-related knowledge has been provided to relevant stakeholders in Europe. All partners contributed to their realisation and success, whether through attendance at conferences and expert meetings, contributions to scientific and popular publications etc.
Publications
Communicating on PREDICT findings in the scientific community with papers published in journal during the project had three main impacts:
▪ Contributes to the overall dissemination efforts of PREDICT;
▪ Ensuring that the findings of the project will be useful for future research and projects;
▪ Validating these findings by peers.
In PREDICT, publications have been deliberately aimed at different journals to access both the scientific community and the practitioners community.
The table below presents some of the publications produced during PREDICT.
Mapping of areas presenting specific risks to firefighters due to buried technical networks Author: Amélie Grangeat (CEA)
Paper submitted to a Special Issue of the International Journal of Information Systems for Crisis Response and Management (IJISCRAM) and submitted to ISCRAM Conference 2016
Human vulnerability mapping facing vital services disruptions for crisis managers Authors: Amélie Grangeat (CEA), Julie Sina, Marianthi Theocharidou, Vittorio Rosato, Aurélia Bony
Paper submitted to the ISCRAM 2016 conference.
Submission of an article on the progress of the project for the December 2016 issue Authors: Olivia Cahuzac and Martin de Maupeou (CEIS)
Presentation of the project, its progress and the next steps. Emphasis on the involvement of crisis management practitioners throughout the project
Addressing Cascading Crises in Europe in Crisis Response Journal 10:2 Authors : Olivia Cahuzac (CEIS)
Presentation of the project, its objectives, with a focus on its added value for crisis management practitioners.
Incident evolution framework for crisis situations – Methodologies for the operation time model Authors: Kling, T. and Hakkarainen T. (VTT)
Abstract and presentation related to WP3, in the Nordic Fire & Safety Days 2016. Aalborg University, Copenhagen, Denmark, 16-17 June 2016.
A method for visualisation of uncertainty and robustness in complex long-term decisions. Authors: Hanski, J. & Rosqvist, T. (VTT)
An abstract submitted to the 2016 European Safety and Reliability Conference (ESREL 2016), Glasgow 25-29 September 2016
Attractor-Directed Particle Filtering with potential Fields Author: Patrick De Oude (Thales NL)
Paper presented at the Fusion 2016 conference (www.fusion2016.org) in Heidelberg, Germany.
Elicitation of a utility from uncertainty equivalent without standard gambles Authors: C. Labreuche, S. Destercke, B. Mayag. (Thales FR)
Paper presented during the 13th European Conference on Symbolic and Quantitative Approaches to Reasoning with Uncertainty 2015 (ECSQARU'15). Compiègne, France.
Dependencies and Chain Effects in National Security & Crisis Management Journal Publications from PREDICT partners Marieke Klaver, Eric Luiijf (TNO), Rob Peters, Nico van Os (VRZHZ) and René Willems (TNO) appear in the "Magazine Nationale Veiligheid en Crisisbeheersing" 2015
Conferences and events
The “Cascading Effects” Joint final conference - Brussels (Belgium), 16 & 17 March 2017
PREDICT’s final dissemination activity is the joint final conference where the key project results will be presented at the Cascading Effects Conference,
The conference also presents and discusses results, models and tools from PREDICT, CascEff, CIPRNet, FORTRESS and SnowBall – all projects funded by the European Union’s Seventh Framework Programme.
DOMINO II Conference – Tiel (The Netherlands), 21 & 22 September 2016
On 21 and 22 September 2016, Thales (NL and Fr), ITTI, TNO, and VRZHZ participated in the DOMINO II Conference also held in Tiel and organised by the Dutch First Responder platform, IFV (http://www.ifv.nl).
Several European project took part in this event which features live demos of information sharing systems and presentations on the issues of interoperability in crisis management.
Over 20 crisis management practitioners from all over the Europe, including operators, trainers and high-level decision makers, had an opportunity to get familiar with crucial iPDT functionalities during 3 hands-on sessions.
DOMINO Research and Development – Zwijndrech (The Netherlands), 21 & 22 May 2015
The Dutch First Responder platform IFV organised a joint research and development event about crisis information management during incidents with cascading effects attended by some of the PREIDCT partners. The event - named ‘DOMINO’ - took place May 20, 21st and 22nd 2015.
Over the last five years, the Dutch platform of government emergency professionals – closely related to the National security council and the joint Fire Brigade Commanders of the Netherlands - has been working on issues of interoperability, semantics, geo-information, interpretation and multi-actor data preparation in order to improve the task they have all in common; to improve the information position of the incident managers during a large crisis. Six European FP7 research projects currently addressing this problem including the FP7 PREDICT project took part in this event.
Research Executive Agency (REA) workshop on synergies between FP7 projects – Brussels (Belgium), 25 June 2014
On the 25 June 2014, representatives from the PREDICT consortium (CEIS) took part in a workshop organised by the Research Executive Agency (REA) to explore synergies between FP7 projects related to critical infrastructure protection against natural hazards and cascading effects in crisis situations.
The aim of this meeting was to establish an effective collaboration between different projects, which are dealing with similar issues (critical protection against natural hazards, extreme events LP-HI, and cascading effects in crisis situations), to avoid redundancies and potential duplication of efforts, and to improve the quality of the expected results and boost their impact.
List of Websites:
Website: www.predict-project.eu CEIS
Coordinator:
Dominique Sérafin CEA (Commissariat à l’énergie atomique et aux energies alternatives)
Gramat Center
BP 80200
46500 Gramat
France
dominique.serafin@cea.fr
+33 5 65 10 54 46