基于自适应进化学习的约束多目标粒子群优化算法

Constrained multi-objective particle swarm optimization algorithm based on self-adaptive evolutionary learning

针对约束边界粒子在边界区域搜索能力不足的问题,提出一种基于自适应进化学习的约束多目标粒子群优化算法.该算法根据不符合约束条件粒子的约束违反程度,修正优化算法的进化学习公式,提高算法在约束边界区域的搜索能力;通过引入一种基于拥挤距离的Pareto最优解分布性动态维护策略,在不增加算法复杂度的前提下改进Pareto前沿的分布性.实验结果表明,所提出的算法可以获得具有更好收敛性、分布性和多样性的Pareto前沿.

Considering the problem of the inadequate search ability for constraint boundary particles in the boundary region, a constrained multi-objective particle swarm optimization algorithm based on self-adaptive evolutionary learning is presented. The evolutionary learning formulas of multi-objective particle swarm optimization algorithm are modified according to the constraint violation level of infeasible particles, so that the algorithm＇s search ability is enhanced greatly in the constraint boundary region. Furthermore, a dynamic distribution maintenance strategy for Pareto front based on the crowding distance is adopted to improve the distribution of Pareto front without any increase in the algorithm＇s complexity. The experimental results show that the Pareto front obtained by the proposed algorithm has better convergence, distribution and diversity.