联盟是多Agent之间一种重要的合作方法,如何生成面向某个任务的最优联盟是一个复杂的组合优化问题。提出一种贪婪算法:逐步选择加入联盟的Agent,直到联盟的能力满足任务要求,此时完整的联盟被生成。在联盟生成过程中,任一Agent加入联盟的贪婪准则是它能否比其他Agent带来更大的能力和更小的成本。实验结果表明本算法是切实可行的。 Coalition is an important cooperative method between multiple agents. How to generate the optimal coalition for a certain task is a complex combinatorial optimization problem. Proposed a greedy algorithm: gradually choose to join the coalition of the Agent, until the coalition’s ability to meet the mission requirements, then a complete coalition was generated. In the process of coalition generation, the greedy criterion that any agent joins the coalition is whether it can bring more capacity and less cost than other Agents. Experimental results show that this algorithm is feasible.