一种防空武器发射装置平衡机优化设计混合算法  

A Hybrid Algorithm Approach to the Optimization Design of Air-Defense Launcher Balancer

作　　者：韩勇[1] 王洁[1] 程永强[1,2] 任卫华[1] 吕爽[3] 

机构地区：[1]空军工程大学防空反导学院,陕西西安710051 [2]空军工程大学防空反导学院,陕西西安710051 [3]中国飞机强度研究所,陕西西安710065

出　　处：《火炮发射与控制学报》2013年第4期39-43,共5页Journal of Gun Launch & Control

基　　金：航空科学基金资助项目(20120196006);空军工程大学研究生创新基金(DY13204)

摘　　要：平衡机是变射角发射装置中的重要组成部分。针对平衡机的优化问题,对平衡机工作原理进行了分析并建立了优化模型,对推式平衡机进行了优化。研究发现:平衡机优化是极大极小值(minimax)问题,考虑到minimax问题的非光滑特性给求解带来了困难,采用极大熵法将问题转化为光滑函数,并利用遗传算法对问题进行求解。采用三点交叉法改善标准遗传算法的变异效果,并利用随机方向法对其结果进行搜索,改善了遗传算法的早熟问题。通过对比发现,混合算法的优化结果分别较单纯形法与复合形法的优化结果改进了41.6%和23.8%,且改进的遗传算法优于标准遗传算法。Balancer is an important component of turning angle launcher. Aiming at the optimal design of balancer, an optimal model was established, an optimal answer was given to the spring push balancer. Studies showed balancer optimization a minimax problem. Considering the fact that nonsmooth characteristics bring difficulties to the minimax problems, maximum entropy method was used to convert it into a smooth function. Then genetic algorithm was used. To overcome the premature problem of GA, three crossing method was used to improve variation effects of the standard GA, and stochastic direction method was used to improve the result. Comparing with the result of simplex method and complex method, the hybrid method improved by 41.6% and 23.8%, and the improved GA is better than the standard GA.

关 键 词：平衡机 优化设计 MINIMAX 极大熵法 遗传算法 

分 类 号：TU323.4[建筑科学—结构工程] TU311.4