Strategic Reorganization in Multi-Agent Systems: Inspired by Intelligent Human Organizations

Abstract

The field of organizational design in multi-agent systems is still in need of powerful tools and methods for enabling effective design, control, and transformation of organizations of different kinds. In this work, we address the problem of reorganization in a distributed problem solving model as a typical multi-agent system application. We look into adaptation of concepts and theories from social organization theory. In particular, we get insights from Schwaninger's model of Intelligent Human Organizations. To this goal, we implement and utilize a multi-level organizational control model which uses a strategic management and an operative management layer to exert organizational control and also to make interactions between these two levels of control possible. Experimental evaluations are performed using a modified model of pursuit game. The results indicate that the proposed model allows the system to stay ahead of the organizational change, resulting in performance improvements, despite the additional costs associated to reorganization.