结合增广Lagrange罚函数的约束优化差分进化算法

利用增广Lagrange罚函数处理问题的约束条件,提出了一种新的约束优化差分进化算法。基于增广Lagrange惩罚函数,将原约束优化问题转换为界约束优化问题。在进化过程中,根据个体的适应度值将种群分为精英种群和普通种群,分别采用不同的变异策略,以平衡算法的全局和局部搜索能力。用10个经典Benchmark问题进行了测试,实验结果表明,该算法能有效地处理不同的约束优化问题。

Constrained optimization differential evolution algorithm usingaugmented Lagrange penalty function

Using augmented Lagrange penalty function to deal with the constrained conditions, this paper proposed a modified constrained optimization differential evolution algorithm. It converted the general constrained optimization problem into a bound constrained optimization problem. In the process of evolution, divided the initial population into two subpopulations, i. e. elite and general subpopulations, which used different mutation strategies to balance the ability of global and local search respectively. It tested ten classic Benchmarks problems, the experiment results show that the proposed algorithm is an effective way for constrained optimization problems.