论文部分内容阅读
为解决约束优化问题,提出使用双可行域吸引子策略改进动量粒子群算法。该算法只需初始种群中有一个粒子位于可行域内,随着搜索过程的进行,整个种群自动进入可行域内搜索。一方面,在搜索过程早期,由于可行域内粒子少,所有粒子移向相同的吸引子,整个种群迅速进入可行域内。另一方面,随着进入可行域粒子的增多,由于每个粒子使用距本身最近的可行域吸引子,较好地维持了种种群的多样性,避免早熟现象的发生,使算法具有较好的寻优性能。与国际上当前解决约束优化问题的粒子群算法在4个标准约束优化函数上测试比较,实验结果表明本算法取得的最优值要优于其它粒子群算法。
In order to solve the constrained optimization problem, this paper proposes a new method of momentum particle swarm optimization using dual feasible domain attractor strategy. The algorithm only needs one particle in the initial population to be located in the feasible region, and as the search process proceeds, the entire population automatically enters the feasible region search. On the one hand, early in the search process, because of the small number of particles in the feasible region, all the particles move to the same attractor, and the whole population quickly enters the feasible region. On the other hand, as more particles enter the feasible domain, each particle uses the feasible domain attractor closest to itself to better maintain the diversity of the species and avoid the premature phenomenon, so that the algorithm has better Search for excellent performance. Compared with the current international PSO algorithms for solving constrained optimization problems, the proposed method is tested on four standard constrained optimization functions. The experimental results show that the optimal value obtained by this algorithm is superior to other PSO algorithms.