Agent-based Simulation Framework for Safety Critical System

Abstract Safety critical Systems often integrate multiple complex functions and must generally ensure safety. These requirements lead to a need for simulation system that provides a prototype system for evaluation. This article proposes an agent-based simulation framework to meet the challenges of designing and evaluating the performance of the safety critical system. The approach proposed represents each agent architecture including Agent communication interface, Reasoning, Modeling, domain knowledge set, Execution and human agent interface units. The agent maps the safety requirement using risk ordering. The whole simulation system composed of multiple agents. One case study shows that agent-based simulation can express the behavior and performance of the system as a whole, especially, safety attributes can be displayed.