多样本遗传算法在武器外弹道组网试验中的应用

Application of Multi-sample Genetic Algorithm in Weapon Exterior Ballistics Networking Test

针对武器试验外弹道组网测量模式,为保证弹道测量精度最优,对测试设备优化布站进行研究。以弹道逐点最小二乘解算模型为基础,弹道参数累积总误差的均方差为精度评价标准,考虑实际情况对解空间构成的非线性约束条件,通过聚焦遗传算法概率特性,设计以MonteCarlo实验和种群罚函数检验为特征的完备多样本遗传算法,实现优化问题的可靠求解。利用设计的遗传算法,对一种光学、雷达设备组网测试优化布站问题作500次MonteCarlo仿真实验,通过结果统计分析确定了全局最优解,并与传统布站方法进行了比较。仿真结果表明:多样本算法降低了遗传算法概率特性影响,以大概率事件锁定全局最优解,确保了最优解的有效性和可信度,符合工程实际应用。

For the networking measurement mode of exterior ballistic in weapon test, the optimal disposition of test equipment is studied to ensure the optimal ballistic measuring precision. Based on the point-by-point least squares solution model, the root-mean-square of total cumulative error of ballistic parameters is designed as precision evaluation criterion. Considering the nonlinear constraints of solution space in actual situation and focusing on the probability characteristics of genetic algorithm, a complete multi-sample genetic algorithm based on Monte Carlo test and population penalty function check is designed to reliably solve the problem of optimal disposition. The proposed genetic algorithm is used to simulate and evaluate an optimal disposition of test equipment. Monte Carlo experiments were made for 500 times, a globally optimal solution was obtained by statistically analyzing the experimental results, and the proposed algorithm was compared with the traditional method. The simulated results show that the proposed algorithm can be used to reduce the influence of probability characteristic of genetic algorithm, lock the global optimal solution with large probability events, and ensure the validity and reliability of the optimal solution.