【作者】 周琴；

【导师】 任光；

【作者基本信息】 大连海事大学 ， 控制理论与控制工程， 2003， 硕士

【摘要】 本文研究了遗传算法在函数优化、生产调度和自动控制三方面的应用，在函数优化方面，详细研究了求多峰函数的最大值问题；在生产调度方面，讨论了作业车间调度问题；在自动控制方面，探讨了PID控制的PID参数寻优问题。 对于求多峰函数的最大值问题，通过对传统优化方法单纯形法和遗传算法各自的特点分析，本文提出了一种结合二者优点的混合遗传算法。该方法利用了单纯形法的局域搜索能力和遗传算法的全局搜索能力，通过遗传算法控制大范围搜索的方向，使得搜索向着适应度值高的区域发展，再通过单纯形法在遗传算法搜索到的区域内进行小范围的邻域搜索，从而能够得到高适应度值域的最优值。本文提出的算法是在遗传算法的交叉和变异算子之后对每个个体进行单纯形局部寻优的单纯形混合遗传算法，并且对变异算子进行了改进，采用大变异算子，以达到加快收敛速度的目的。详细阐述了混合遗传算法的设计、实现方法，给出了采用MATLAB编程的程序框图。通过仿真研究和与其他方法的比较，该大变异单纯形混合遗传算法在收敛精度、收敛代数、收敛率方面都有所提高，具有一定的研究价值。 对于作业车间调度问题，本文分别探讨了基于工序编码和基于工件编码的调度问题的算子选择和算法实现，对于关键算法如适应度评价函数的实现进行了详细的讨论，并给出了具体的实现步骤。 船舶PID控制的PID参数寻优问题。以评价系统快速响应性能的ITAE为标准，采用其转化形式为评价PID参数适应度的适应度函数，在PID参数的允许范围内，通过仿真得到ITAE值，从而得到PID参数的适应度值。再通过遗传算法的搜索，得到参数允许范围内的一个较好的解。更多还原

【Abstract】 In this paper we investigate the application problem of genetic algorithm to function optimization, production precedence and automation. We study multi-maxim function in the domain of function optimization, Job-shop in the domain of production precedence and the PID parameters selecting in the domain of automation.For solving multi-maxim function, we comes up with a hybrid genetic algorithm through analyzing characteristics of pattern search and genetic algorithm. This hybrid algorithm take advange of local search of pattern search and whole search of genetic algorithm. It first confines the direction of large scale search to the area of high fitness by genetic algorithm, and then utilizes pattern search to search in the small area got by the searching of genetic algorithm., thereby the optim value is got. This hybrid algorithm applicates pattern search on each individual afer crossover operator and mutation operator of genetic algorithm, and improve the mutation operator to the big mutation operator in order to speeding convergence. The paper expatiate on the design and realizing of the hybrid algorithm, and present the flow chart of the program programmed with MATLAB. Through studying the simulation results and compared with the other algorithms, the big mutation hybrid algorithm improves convergence accuracy and convergence probability and reduce convergence generation, therefore it is worth further researching.For the job-shop problem, the operator selecting and algorithm realization based on working procedure coding and based on job coding is investigated in the paper. Some key algorithm such as fitness evaluation function is given in detail.Genetic algorithm is employed to search for the optimized parameters of PID controller applied to ship controling. According to ITAE criterion which appraises system response performance, the value of ITAE is got by simulation of system in the allowable range of PID parameters, and then the transformation of the value of ITAE is adopted to be fitness function evaluating fitness of PID parameters. Better solutions in the allowable range of parameters are gained through searching by genetic algorithm.更多还原