变粒度协同进化设计算法及其在卫星舱布局设计中应用

采用进化算法求解复杂卫星舱布局问题时,算法容易陷入局部最优,且干涉计算复杂度高,计算耗时长.为提高对复杂解空间的搜索能力,基于协同进化算法,将问题分解为若干子问题求解;为减少计算耗时,子问题求解时采用了一种设计变量的变粒度策略.称上述方法为变粒度合作式协同进化算法(CCEA-CFG).卫星舱布局优化数值实验表明,与目前常用的几种布局求解算法(遗传算法、协同进化算法以及遗传/粒子群算法(QPGP))相比,CCGA-CFG(基于GA的CCEA-CFG)具有较好的计算质量、计算效率和计算鲁棒性.

Coevolutionary algorithm with coarse-to-fine grain strategy and its application to layout design of satellite module

The main problems for evolutionary algorithm to deal with the complex layout design of satellite module include easily being trapped into local optimality,high complexity of interference computation and large amount of consuming time.To enhance the global search ability in the complex solution space,the layout design problem was decomposed and cooperative coevolutionary algorithm was employed for optimization search.Meanwhile,in order to reduce the computational time,a coarse-to-fine grain strategy of desig...