An overview of a multiagent-based simulation system for dynamic management of risk related to Dangerous Goods Transport

Transportation engineering, and especially of Dangerous Goods (DG), is one of the areas in which technological development is crucial. Logistic control systems, routing and scheduling algorithms, supply chain management have been applied in this aspect so as to improve the cost-effectiveness ratio and to reduce risks. Unfortunately, most of these works remain theoretical since tests on real case scenario sometimes proves difficult to achieve especially when we are dealing with Dangerous Goods Transport (DGT). In this paper, an architecture of a visual DGT simulation system has been introduced. It provides entries for testing these works and produces results close to real experiments. Applications on serious games, financial market, engineering are evidence of success of the visual simulation system. Thus, the challenge consists on how to design a system that represents real-world systems with an appropriate degree of complexity and dynamics. Several researchers have already suggested solutions since the 50s. These solutions have been classified in three levels: microscopic, mesoscopic and macroscopic simulation. It has been decided to use a microscopic approach using an agent-based model. It allows the representation of systems at different levels of complexity, a System of Systems (SoS), through the use of autonomous, goal-driven and interacting entities. The system is implemented using the software MATSim (http://www.matsim.org/). The main actors of the system stand for the DG Carriers Operator (CO) agent that manages a fleet of DG Carriers agents, taking logistic decisions on their behalf. The COs collaborate with the National Authority (NA), which is an agent concerned by the safety shipment of goods at regional and national level. The paper focus on the Multi-Agent System (MAS) architecture and introduces its key components as well as the means on interaction.