基于模拟退火算法的生产调度现象的研究
摘 要
公司在实际运作过程中，其自动化方面的轴心维度主要是生产管理和生产调度。从实际效益来看，科学有效的调度对策可以使得公司获得极为可观的经济实效。就理论层面而言，生产调度也并不是单一化的，是数个目标及其约束条件下的优化集成问题，因此，研究该课题显然具有突出的理论性与实用意义。
生产调度并不简单，其涉及到的知识范围不一而足，而其生产环境也处于同步共时的状态之中。这些特征决定了研究该问题是无法仅仅凭借人工抑或是电脑即可解决的。除却人力、计算机工具之外，还要考虑到必要的数学规划知识，并将上述几点整合在一个体系之中，才可以有效地解决有关生产调度的课题。
在研究工具方面，本论文所主要采用的是模拟退火算法。该算法是一类比较新颖的方式，具有突出的随机性与搜索功能。它的问世与最近几年来所需要的针对规模不小的组合优化课题密切相关。它并不是精确型数学，而是偏于近似化的算法。和以前的近似法相较而言，该法比较便捷灵巧，使用的范围比较地广、运作效率较高以及少到最初条件制约性不大等优势，尤其适用于并行化的演算。它的优化功效包括一系列函数关系式，主要有状态产生及其接纳以及温度更置等。同时，该算法还适合于内、外循环的相关定则，凸显出该算法的迅捷化、收敛化以及鲁棒化等特点。
具体而言，本论文的研究内容分析安排如下：
第一章是绪论。对选题背景及其意义加以简介，系统化梳理了国内外在生产调度以及模拟退火算法方面的研究现状，再简单地介绍了本论文的研究内容及其研究方法。
第二章概述了模拟退火算法理论。首先对VFSA理论加以概述，包括它的模型扰动情况、接受概率以及退火计划3个方面。再对VFSA内在机理展开了相关的研究。在前面分析的基础上提出了改进的模拟退火算法。
第三章系统化地构建起生产调度优化模型。首先分析的是从生产调度模型的分析、它的约束条件及其成本模型3个方面来构建相应的模型。其次，探讨有关生产调度模型的求解方法。
第四章是优化生产调度过程实例描述。在比较全方位地提出相应的问题之
后，再依次条理化地开展编码、算法原理、交叉、变异与选择等步骤的分析，并概要了有关并行模拟退火遗传算法流程。最后是从物理退火过程、Metropolis准则以及模拟退火算法3个方面来对模拟退火算法展开比较详细的描述。
第五章分析的是优化生产调度的实现及其仿真。首先分析如何确定解空间，然后分析怎样挑选初始解，再次分析新解的产生及其接受机理情况，接着分析的是补充停止准则，最后分析的是仿真结果。
第六章是全文的结论。
关键词：模拟退火算法，生产调度，退火计划，最优，数学模型
Abstract
In the actual operation of the process, the automation of the axial dimension is the main production management and production scheduling. From the practical benefits of view, scientific and effective scheduling strategy can enable the company to get considerable economic efficiency. As for the theory, production scheduling is not single, is to optimize the integration problem, several objectives and constraints. Therefore, this research has theoretical and practical significance.
Production scheduling is not simple, the scope of knowledge This is not the only one. involved, and the production environment is in synchronous synchronic state. These characteristics determine the problem is not only rely on the manual or computer can solve. Beyond the human, computer tools, but also consider the mathematical programming knowledge is necessary, and a few of the integration in a system, it can effectively solve the problem of production scheduling.
The research tools, the main use of this thesis is the simulated annealing algorithm. The algorithm is a kind of novel style, with random and search features prominently in the. It is the advent of recent years needed for combinatorial optimization problem is not a small scale is closely related to the. It is not the exact mathematics, but partial to the approximation algorithm. And compared with the approximation method used, the method is convenient and smart, use range is broad, higher operating efficiency and less to the initial conditions of small advantages, particularly suitable for parallel algorithm. Optimization includes a series of relationship, mainly produced and its acceptance as well as the temperature set etc.. At the same time, the algorithm is also suitable for internal, external circulation related rule, highlights the faster convergence and robustness, so the algorithm.
Specifically, analysis the research contents of this thesis are as follows:
The first chapter is introduction. On the background and significance of topics to be brief, systematic reviews the research status at home and abroad, production scheduling and simulated annealing algorithm, and then simply introduces the
research content and the research method of this dissertation.
The second chapter outlines the theory of simulated annealing algorithm. The VFSA theories, including its model, perturbation probability of acceptance and annealing plans 3 aspects. Then the internal mechanism of VFSA is studied. On the basis of above analysis, put forward the improved simulated annealing algorithm.
The third chapter constructs the optimization model of production scheduling. The first analysis is from the analysis, production scheduling constraints and its cost model 3 aspects to construct the corresponding model. Secondly, to explore the solution on the production scheduling model.
The fourth chapter is the optimization of production scheduling process instance description. After the corresponding problem is full, then organized to carry out coding, algorithm, cross, variation and selection steps of analysis, and summary about the parallel process simulation annealing genetic algorithm. Finally, from 3 aspects of physical annealing process, the Metropolis criterion and the simulated annealing algorithm to launch a more detailed description of the simulated annealing algorithm.
The fifth chapter is the analysis of realization and simulation and optimization of production scheduling. Firstly, how to determine the solution space, and then analyze how to choose the initial solution, again a generation and accept mechanism analysis, then the analysis is the supplement stopping criterion, finally is the analysis of the simulation results.
The sixth chapter is the conclusion of this paper.
Key words: simulated annealing algorithm, production scheduling, annealing scheme, the optimal, mathematical model