CONCEPTUAL FRAMEWORK FOR EVALUATION OF THE EFFECTIVENESS OF INTELLIGENT SECURITY MEASURES IN PUBLIC TRANSPORTATION THROUGH A MULTI-AGENT-SIMULATION DATA FARMING EXPERIMENT TO PREVENT TERRORIST ATTACKS

摘要

Dealing with terrorist attacks is a major challenge - especially for critical infrastructure protection when hazards cannot be described in normative risk management systems, e.g., anthropogenic threats with low probability but high consequences. Using the example of rail-bound public transportation, a multi-agent simulation of a terrorist attack on a fictitious train station provides a basis for evaluating the effectiveness of intelligent security measures. Using different scenario constellations, the challenge of assessing low terror event probabilities empirically or historically can be replaced by conducting simulation experiments for a wide bandwidth of possibilities and analysing this data for further risk management assessment. This process is called Data Farming: in this paper, it is applied to a specific case and the results obtained are discussed. It is shown that the proposed approach gives essential support to decision-makers who are in charge of shaping security systems according to the given requirements. Furthermore, the conceptual framework includes an expert table-top-exercise and a realistic exercise in a real train station in Cologne. The outcome and the course of the exercise were captured numerically to compare it with the simulation results and used to calibrate the simulation model.