多目标联盟竞赛算法
Multiobjective league championship algorithm

提出一种基于联盟竞赛的多目标进化算法，根据Pareto占优机制重新定义原算法在团队阵型比赛输赢上的判断，并优解扩散策略使算法不轻易陷入局部最优解，最后结合优解扩散策略开关和分布性指标定义算法终止条件. 通过对4个二维目标ZDT测试函数、4个三维目标DTLZ测试函数的实验及其他多目标进化算法的对比和分析，验证了新算法的可行性和有效性.
A multiobjective evolutionary algorithm based on league championship is proposed in this paper. Firstly according to pareto dominant mechanism redefine the judgment of team winning or losing. Secondly proposing optimal solution diffusion strategies avoid to fall into local optimal solution.Lastly using optimal solution diffusion strategies switch and spacing metric enable the algorithm auto stop.The proposed algorithm and other two algorithm are applied to four ZDT test functions and four DTLZ test functions.Experimental result show that the new algorithm is feasible and efficient