应用遗传算法设计自动交会控制器

本文提出应用遗传算法求解黎卡及方程，并将这一方法用于设计自动交会控制器．遗传算法是一种基于自然选择和群体遗传学机理的参数搜索方法，它将达尔文的适者生存策略与人工染色体中的随机而结构化的信息交换相结合，既消除了求解过程的不适应因素，又利用了解中已有的知识，从而大大加快了优化速度和能力，且计算比直接解黎卡提方程要简单．航天器自动交会控制系统的仿真结果表明，这一方法是有效的．

Design of Automatic Rendezvous Controller Using Genetic Algorithm

The authors make use of the recently revived genetic algorithm(GA)method to design LQ (LinearQuadratic)regulator so as to improve greatly its ability to solve Riccati equation in the problem ofautomatic rendevous and docking(RVD)system.GA reduces greatly computation, and guarantees rapal and ace convergenceIn the GA dcsigD of regulator for RVD.the quadratic performance index is rcgnrded as the fitnessfunction and the feedback gain matrix is obtained mathematically by utiliaing three genetic operators:sckction(reproduction),crossover(mating),and mutation. Simulation resultS allow that,for solvingXieeati equation, GA method bas four advantages,namely.I)Fast convergence.GA,which is called implicit parallelism,searches from a population ofPOints, not just a single point. From Fig.2 the results converge after only 40 generations.2)TrUe global optimum, not local optimum. GA uses probabilistic transition rules,notdeterministic transition rules3) Simplicity in computation. Each gain in the gain matrix is represented by a 8-bit string.Thestrings are concatenated to produce 64-bit string,each string representing one feedback design.GAworks with these codings(strings),not the feedback gains themselves4) Wide applicability.GA considers the quatratic performance index as the fitness function.GA"scsit,Instead of derivative or other information,to evaluate potential solutions.GA can operate'blind.