基于均匀设计与RBF网络的火炮优化目标函数构建方法研究  

Research on establishment measures of artillery objective function by uniform design method and RBF network

作　　者：崔凯波[1] 秦俊奇[1] 狄长春[1] 姚世锋 

机构地区：[1]军械工程学院火炮工程系,河北石家庄050003 [2]重庆军代局,重庆400054

出　　处：《机械设计》2013年第2期45-48,共4页Journal of Machine Design

基　　金：国家自然科学基金资助项目(50645019)

摘　　要：为解决虚拟样机优化运算量大、耗费时间较长等问题,达到脱离虚拟样机仿真环境、提高参数优化效率的目的,以炮口扰动最小化为优化目标,选取待优化结构参数,利用均匀试验设计方法和虚拟样机仿真结果建立输入输出样本库,用基于K-means聚类算法的广义RBF网络求解得到待优化结构参数与炮口扰动指标的非线性映射关系,并对相应指标进行加权处理,最终确定了炮口扰动优化目标函数,为有效利用优化算法对火炮结构参数进行优化提供了依据。In order to solve the problems such as complicated computation and long work time, achieve the objective of simulation environment without virtual prototype and improving the parameters optimization efficiency, by taking muzzle disturbance minimization as optimization objective, the structural parameters to be optimized were chosen. The input/output sample database was set up based on the uniform design method and simulation results of virtual proto- type. Use generalized RBF network based on K - means clustering algorithm to solve and obtain the nonlinear mapping relationship be- tween structural parameters to be optimized and muzzle disturbance index, and the corresponding indexes were weighted. The muzzle optimization objective function was established ultimately, having provided a reference for structural parameters optimization of artil- lery by using optimization algorithm effectively.

关 键 词：火炮 虚拟样机 均匀设计 RBF网络 目标函数 

分 类 号：TJ303[兵器科学与技术—火炮、自动武器与弹药工程]