用并行遗传/Powell/蚁群混合算法求解卫星舱布局问题

卫星舱布局优化设计问题数学上属NP-hard问题．较有效的求解途径之一是研究混合算法，为此构造了并行混合PGA-Powell-蚁群算法（简称PGPAHA）．该算法以并行遗传算法为基本框架，根据各子群体收敛速率的快慢来决定它们之间迁移的时机，在收敛后期加入Powell法作为并行遗传算法的算子来加快收敛速度，并利用蚁群优化算法提高计算的精度．最后应用该算法求解了简化的三维带性能约束的国际商用通讯卫星的卫星舱布局设计问题，数值实验结果表明，该算法与并行遗传算法（PGA）相比，在计算精度、计算效率及计算稳定性方面较优。

Layout optimization of a satellite module using a parallel genetic-Powell-ant colony hybrid algorithm

The problem of the layout design of a satellite module(LDSM) belongs to NP-hard problem in mathematics;one effective way to solve this problem is to explore hybrid evolutionary algorithms.Based on the framework of parallel genetic algorithm(PGA),a parallel genetic-Powell-ant colony hybrid algorithm(PGPAHA) is given,in which the migration moment of sub-populations is decided according to their convergence rate,and Powell method is taken as one of operator of PGA to accelerate their convergence rate at the later period of the searching,then ant colony optimization(ACO) algorithm is used to enhance the computational accuracy.The given algorithm is used to solve optimal layout design of a simplified international commercial communication satellite module.The numerical experiment results show that the given algorithm is superior to the PGA in computational precision,efficiency and stability.