A case study of task-based reorganization in a pursuit game simulation

摘要

Important to the effective performance of a multi-agent system, is using the appropriate organizational structure. Different organizations might be appropriate for different tasks, goals, and environmental conditions. As a result, sometimes a higher performance can be achieved by reorganizing to a more promising structure. One of the problems in reorganization is defining the criteria upon which to evaluate and find a promising organization to reorganize to. Considering the costs that can be associated to reorganization, most of the research to-date has been focused on reorganizing when system performance falls below a certain threshold. While some opportunities for reorganization might be lost. In this work, we perform a simulation study to identify effects of a task-based model of reorganization. The results demonstrate performance enhancements using our approach despite the costs associated to reorganization.