Join coalition mechanism for enhancing idle agents’ utility in a macroscopic coalition
摘要
The agents can only maximize their utility under certain formal definitions in the coalition formation (CF) which usually require finding optimal coalition structure. By finding the optimal coalition structure, the profit of the coalition is secured and the agents will cooperate to solve the problem using distributed approach. However, the idle agents are not able to maximize their utility because neither they have insufficient budget to form a coalition nor receiving invitation from others. The join coalition mechanism (JCM) has been introduced for providing idle agents an opportunity to join an existing macroscopic coalition for maximizing their utility. The JCM uses two phases verification which are voting session of the agents’ offer (namely join coalition request) and trial joining session using n-players iterated prisoner dilemma (NIPD) games. The experiments study the behavior of agents in verification phases of the JCM. The experimental shows that the agents are able to conserve their communication rate between up to 88% in a high defective ratio coalition. Besides, the agent with flexible cooperative behavior has a higher chance of joining coalition. Through the simulations, we have observed that the communication has tune down and the idle agents are able to join the targeted coalition if the mutual trust has been developed.