基于序列近似优化算法的固体运载火箭弹道设计

采用打靶法研究固体运载火箭弹道优化设计问题。针对打靶法中智能优化算法效率、优化问题中约束条件提法等方面的不足,引入序列近似优化算法,提出算法中径向基函数高斯核宽度的高效确定方法。数值仿真实验表明,提出的方法可在基本不带来计算量增加的前提下,显著提高代理模型精度,使序列近似优化算法得到改进,提高优化效率。设计固体运载火箭飞行程序。建立弹道优化设计问题数学模型,并将模型中相关等式约束合理转化为不等式约束,降低优化问题求解难度;基于改进后的序列近似优化算法完成某固体运载火箭弹道优化,原始计算模型调用308次之后便搜索到最优解,较传统智能优化算法显著提高了优化效率,优化方案末助推级液体推进剂消耗比原方案减少25.7%。

Solid launch vehicles trajectory design based on sequential approximate optimization algorithm

Solid launch vehicles trajectory optimization problem was researched using shooting method.Current shooting methods have shortcomings on the efficiency of intelligent optimization algorithms and constraints formulation.The sequential approximate optimization (SAO) algorithm was introduced to solve solid launch vehicles trajectory optimization problem.A high efficiency Gaussian kernel width determining method for radial basis function (RBF) used in SAO was proposed.Numerical experiments show that the proposed kernel width determining method improves the approximate accuracy of RBF remarkably and generates almost no excessive calculation cost.Flight program and trajectory optimization problem for solid launch vehicles were established.In order to reduce the difficulty of solving solid launch vehicles trajectory optimization problem,equality constraints in optimization problem were transformed into an inequality constraint ingeniously.Based on improved SAO algorithm,trajectory optimization for a launch vehicle was accomplished,and the global optimal solution was obtained after calculating original model 308 times.Propellant consumption of the upstage in optimization scheme is 25.7％ less than primary scheme.