Multi-Agent Resource Distribution Control with Soccer Agents

In most of the multi-agent systems, resource distribution is one of the major applications. An efficient distribution control can further improve the results and utilize the resources among agents. In this paper, we exploit multisoccer agents to tackle with a human resource management system. By consideration of the soccer tactics and formation, it can distribute the resources more evenly. In addition, various local constraints are appended to evaluate the system. It shows that soccer agents are still able to produce a better allocation plan compared with various brute-force and heuristic search algorithms. Moreover, we describe the implementation issue of both centralized and decentralized control as the view of coach and player approach of soccer agents. We also introduce a new concept of virtual synchronized clock in the distributed control in order to speed up the decision time without negotiations among agents.