基于混合遗传算法的时间最优控制问题求解  

Time-Optimal Control Based on Hybrid Genetic Algorithm

作　　者：赵瑞艳[1] 李树荣[1] 

机构地区：[1]中国石油大学(华东)信息与控制工程学院,山东东营257061

出　　处：《控制工程》2011年第3期417-419,共3页Control Engineering of China

基　　金：国家自然科学基金项目(60974039);国家科技重大专项"高温高盐油田提高采收率技术"课题(20082X0501)

摘　　要：研究了用混合遗传算法求解时间最优控制问题。混合遗传算法是用粒子群位移转移的思想改变遗传算法的变异规则,通过记录各染色体的历史最优值和种群的最优值,来修正下一代的染色体,新的算法保留了遗传算法的选择和交叉操作,保证了遗传算法强大的全局搜索性能,该算法可求解数学优化问题。在分析时间最优控制问题已有求解方法优缺点的基础上,提出基于混合遗传算法求解时间最优控制问题的直接方法,为了利用该算法求解时间最优控制问题,将约束作为惩罚项包括在目标函数中,以此构造适应度函数。对线性阻尼振子问题进行了数值仿真,仿真实例验证了该算法的有效性。The solving of constrained time optimal control problems is studied by using a hybrid genetic algorithm （HGA）. HGA is a kind of GA in which the position displacement idea of the particle swarm optimization（PSO） is applied to change the mutation operation rule, and fix the next generation of chromosomes by recording the history optimal value of each chromosome and the optimal value of population. In this way, the new algorithm retains the selection and crossover operations to ensure the powerful global search performance of the genetic algorithm. The HGA can be applied to solve mathematical optimization problems. A direct method for solving time optimal control problems is proposed on the basis of an analysis of the advantages and disadvantages of the existing algorithms. To realize it in solving constrained time optimal control problems, one construct the fitness function by including the constraints in the objective function as penalty terms. A numerical simulation for the problem of linear damped oscillator shows the effectiveness of the proposed method.

关 键 词：混合遗传算法 时间最优控制 罚函数法 

分 类 号：TP27[自动化与计算机技术—检测技术与自动化装置]