A Layered Multiagent Core for a Decision Support System

Abstract

Decision Support Systems are powerful tools to help support making decisions. However, they are known to be customized for a specific purpose and can rarely be reused. Moreover, they do not support complex situations sufficiently. Our work addresses this challenge and consists in building a DSS that aims to help emergency managers to manage cases of crisis. The DSS is designed to be flexible and adaptive, so that it may be applied on different subjects of studies and whose behaviour may change with the change of its environment. We endowed it therefore with a multiagent layered core whose role is to represent dynamically and in real time the current situation, to characterize it and to compare it with past known scenarios. The final result of the DSS will help decision-makers to analyse the current crisis and its possible evolution. The RoboCupRescue Simulation System is chosen as a test bed to illustrate and to test this approach. Keywords: decision support system; dbscan; clusters; factual agents; factual semantic features; multiagent system