A multi-agent model based on market competition for task allocation: a game theory approach

This paper is devoted to the problem of task allocation in multi-agent systems. Multi-agent systems form a particular type of distributed artificial intelligence system. This paper presents a model based on market competition to solve task allocation problem in MAS. In addition, two algorithms are described in detail which generate reasonable solutions to the task allocation problem based on the following criteria: 1) each agent try to maximize its own profits, and, 2) based on the first criterion, agents try to contribute to the group profits. We utilize game theory to analyze problems of conflict among interacting decision agents. In view of the complexity of calculating Nash equilibria points, one-step Nash equilibrium approach is adopted in the algorithms. Experimental results prove the rationality of the MAS model and the effectiveness of the algorithms presented.