中文摘要：

现有的联盟生成方案多针对一个agent只能加入一个联盟，不利于联盟总效用的最大化以及联盟中agent青色力的充分利用。提出了基于能力类别的agent分解策略，通过定义子agent使得agent可以同时加入多个联盟，在此基础上设计了基于二维离散粒子群的多任务串行联盟生成算法，并对粒子的惯性权重进行动态自适应调整；最后通过算例验证了该方法的有效性。

英文摘要：

Existing researches concentrated mainly on each agent could only participate in one coalition, which goes against optimization of coalition utility and making full use of agents＇ ability. Proposed an agent division strategy according to the abili- ty category and defined subagent to ensure one agent had the possibility of joining in several different coalitions simultaneously. Then designed a two-dimension particle swarm optimization algorithm （PSO） with adaptive dynamic weight value to solve the serial multi-task coalition formation. Finally, proved the effectiveness of the proposed algorithm by the experiment.