一种基于最优算法的导弹鲁棒稳定性评估方法

Evaluation Method of Optimal-based Missile Robust Stability

在对参数不确定性导弹控制系统进行多通道鲁棒稳定性评估时,传统方法在计算效率及结果可靠性方面存在不足。提出一种遗传-Nelder-Mead单纯形混合优化算法,利用遗传算法的全局搜索能力,寻找目标函数的近似最优解,并将其作为Nelder-Mead单纯形法搜索的初值,利用Nelder-Mead单纯形法局部寻优的优势,在近似解的邻域内搜索精确解。通过测试函数验证,算法在计算效率和计算精度方面都有很大提升。最后讨论了算法在导弹控制系统鲁棒稳定性评估中的应用价值。

When evaluating multi-channel robust stability of missile control system with parameters uncer- tainty, the traditional method has disadvantage in computational efficiency and reliability of results. There- fore, a Genetic-Nelder-Mead simplex hybrid optimization algorithm is presented in this paper. Using the global search ability of genetic algorithm to find the approximate optimal solution of the objective function, then we use it as the initial value of Nelder-Mead simplex method ,which has merit in local optimization, to search exact solution in the adjacent domain of approximate solution. It＇ s validated by test functions that the algorithm is improved in computational efficiency and precision of calculation. Finally application of the algorithm in missile control system robust stability evaluation is discussed.