一种新的限制精英的多目标进化算法

A New Multi-objective Evolutionary Algorithm Based on Limited Elitist

在NSGA-Ⅱ算法的基础上,文中提出了一种新的限制精英的多目标进化算法(LEMOEA)。通过分布函数的引入,限制了精英选取的数量,增大了解的搜索区域,从而更好地维护了种群多样性。动态变异算子的引入,减缓了算法的收敛速度,增大了解的搜索区域,避免了算法早熟收敛或陷入局部最优。实验结果表明:LEMOEA比NSGA-Ⅱ有更好的收敛效果和种群多样性。

This paper proposes a new mLdti-objective evolutionary algorithm based on limited elitist （LEMOEA） which is based on NSGA-II. It uses the distribution function to limit the number of individuals chosen by the elitist scheme and increase the ability to search solutions. Therefore, a good diversity of the solutions can be kept effectively. Moreover, the adoption of the dynamic mutation operator reduces the convergence speed of the algorithm, enlarges the area of solution searching and avoids the premature convergence and local optimization of fhe algorithm. Experimental resuits show that LEMOEA has greater convergent speed and better diversity of solutions than NSGA-II.