基于组合优化算法的临近空间飞行器轨迹优化

提出一种临近空间飞行器轨迹优化方法,利用基于支持向量机与遗传算法的组合优化算法,解决多约束条件下的高效轨迹优化问题。首先,建立临近空间飞行器轨迹优化数学模型。然后,通过参数化方法和惩罚函数法将轨迹优化问题转化为约束参数优化问题。在此基础上,提出一种求解无约束参数优化问题的组合优化算法,通过支持向量机对遗传过程中产生的种群进行分类,提高基本遗传算法的计算效率,结合轨迹优化数学模型,给出轨迹优化算法。最后,以临近空间飞行器航程最远轨迹优化问题为例,进行数学仿真分析。仿真结果表明,针对给定的算例,文中提出的方法与基于基本遗传算法的轨迹优化方法相比,计算效率显著提高。

Near Space Vehicle Trajectory Optimization Approach Based on Hybrid SVM and GA Algorithm

A high efficient hybrid algorithm based on support vector machine(SVM) and genetic algorithm(GA) is proposed for near space vehicle trajectory optimization with multiple constraints.The mathematical model for trajectory optimization is derived and the trajectory optimization problem is converted into a parameter optimization problem through parameterization and penalty function approach.The parameter optimization problem is solved by a hybrid trajectory optimization algorithm,in which SVM is used to classify chromosomes generated during GA operation in order to eliminate unfitted chromosomes and accelerate optimization speed An example used for downrange optimization in the proposed algorithm is presented,and simulation results show that the proposed approach is more efficient than the approach based on basic GA.