基于改进遗传算法的调度问题研究-模式识别与智能系统专业论文.docxVIP

摘 摘 要 遗传算法作为一种新型优化算法，由于具有简单、易操作、并行信息处理等 特点，已经在许多领域的优化问题求解方面取得了成功的应用。但是遗传算法在 理论上还不够完善，例如存在容易产生早熟现象以及局部寻优能力较差等问题， 影响了其进一步的应用。本文针对常规遗传算法的不足，提出利用分布种群遗传 算法求解Job-shop和Flow-shop问题，提出了一种基于启发式的遗传算法来求解 单机加工时间可控问题，并进行数字计算研究。 本文主要内容包括以下几个方面： 1．针对常规遗传算法的不足，给出了分布种群遗传算法的抽象形式，并分析 了其收敛性，将该算法运用子Flow-shop和Job-shop问题计算，结果表明 了该算法的有效性。 2．针对单机加工时间可控调度问题，使用基于启发式遗传算法求解单机加工 时间可控问题和单机加工时间离散问题。对具有多变量、非线性和不确定 的此类问题的计算结果表明优化整定后的算法在性能上有了明显的提高· 3．使用基于分布种群的遗传算法对Job-shop问题的加工时间可控问题进行了 研究，完成了对加工可控Job-shop问题的智能算法求解研究，计算结果表 明了所设计的算法具有优良的品质。 关键词：遗传算法，分布种群，NP-hard，单机加工时间可控，Job-shop，Flow-shop Ⅱ AbstractAs Abstract As a new optimization method，GA was widely used in the opfmaizations of manyfieldsowingtothefeatttresofsimpfic毋，easilyhandingandparallelprocessing． However GA theory is not perfect,such as there exist the problems of easily creating earliness and bad ability in local optimal,etc．Enlightened by dismbutien of creature living缸nature，the mathematic model ofthe Distr／bufion Popu／at／on based G∞etic Algorithm(DPGA)is proposed in this paper,and its convergence analysis a∞also given．DPGA is applied to optimize the Job-shop problem(JSP)and How-shop problems(FSP),and the modified GA based Oll hemistic mles is presented for the single machine scheduling problems with controllable processing times．The three problems are typically Nl-hard,which lnealls that it is impossible to find the global optimum in polynomial complexity．Good algorithms for this problem can promote productivity of enterprises．The simulation tests搬made and the results demonstrate the efficiency of the above methedr-1nbe main content of this thesis includes the following： 1．Enlightened by disU-ibution of creature living in natural ecology environment， the mathematic model of the Distribution Population based Genetic Algorithm, and the convergence analysis ofDPGA are given． 2．The algorithms based on DPGA is designed for the JSP and FSP．The simulation tests for some benchmarks show the efficiency ofthe DPGA． 3．The modified GA is presented to