Scenario Generation with Cellular Automaton for Game-Based Crisis Simulation System

Abstract

Crisis is infrequent and unpredictable event which is challenging to prepare and resolve. Scenario generation for modeling a crisis using computational approach is preferable due to its flexibility to produce automated variation of scenario based on different setup key factors. Combining crisis generation with game technology, serious-game game can be developed to provide potential support in training and simulation practice of real-world crisis situation to different stakeholders. Therefore, in this paper, we propose a computational framework to provide automated generation of a fire crisis scenario with firefighter-agent simulation using cellular automaton with influence map for agent behavior optimization, and visualization using a 3D game engine. The framework is evaluated based on performance, convergence and realistic behavior of both crisis event (e.g. fire propagation or control) and crisis resolution (e.g. firefighter solving the crisis).