【作者】 范丹丹；

【导师】 方建安； 王直杰；

【作者基本信息】 东华大学 ， 模式识别与智能系统， 2005， 硕士

【摘要】 服装企业生产流程复杂,目前其自动化程度不高,在实际的生产管理中主要依靠经验丰富的管理人员和调度人员。而且中小型服装企业由于定单多,批量小,其生产计划难于编制,一般都是手工编制生产计划。传统的手工编制生产计划存在效率低、准确度不高、易延误交货期等缺陷。为了适应快节奏的现代化生产和激烈的市场竞争的需要,本文将生产计划智能调度应用到服装ERP中。针对服装企业的生产计划问题,本文建立其数学模型,采用遗传算法进行求解,并且改进遗传算法,最后具体实现了服装ERP中的生产计划智能调度子系统。 首先,本文系统的研究了运用遗传算法求解Flow Shop调度问题的方法,详细讨论了Flow Shop调度问题的遗传算法求解的相关技术,包括编码方法、适应度函数、算法参数、初始化、选择、交叉和变异等遗传操作的设计、算法的终止条件。给出了用遗传算法求解一类典型Flow Shop调度问题的实例。研究结果表明Flow Shop调度问题的遗传算法求解方法具有编码容易、遗传操作简单、寻优性能好的特点。 其次,研究了混合Flow Shop调度问题,并提出了一种改进的遗传算法——多对染色体遗传算法。给出了这一改进算法的新的编码方法和相应的新的个体初始化方法以及交叉操作等。为了进行比较,分别用简单遗传算法和多对染色体遗传算法对同一个具体的混合Flow Shop调度问题进行求解,对这两类算法求解结果的比较表明了多对染色体遗传算法的有效性和优越性。 再次,针对服装企业的生产特点,将服装企业的生产计划问题简化为类似于Flow Shop调度问题的一类问题,给出其数学模型,在此基础上研究了其遗传算法求解方法以及面向定单变化的动态调度方法,计算实例表明该模型和求解方法非常有效。 接着,在我们课题组开发的服装ERP系统中,具体实现了生产计划的智能调度子系统,其中包括:生产计划自动生成、用料单明细自动生成、染整计划自动排缸等功能。它以客户的定单、生产线的生产能力等为依据,合理的安排生产作业计划,对企业的经营决策起到了指导作用,优化了企业的生产运行,取得了较好的效果,而且服装ERP系统包括该子系统已经在道尔奋—帆伦服装公司成功的投入使用。 最后,对本文的主要工作进行了总结,并对整个课题从模型、算法、服装ERP更多还原

【Abstract】 The production line in the clothing industry is so complicated that the management of the production is dependent on the experienced managers who schedule all the procedures. Furthermore, in most clothing enterprises the orders are excessive, so the production planning is difficult to achieve. The production planning that is traditionally made by hand has many drawbacks such as inefficiency, inaccuracy and putting off the delivery date. In order to satisfy the demand of modern rapid production and become more competitive, the clothing industry needs an ERP system attached to the intelligent scheduling subsystem of the production planning. After the mathematic description of a certain planning problem of the clothing enterprise is given, we use the genetic algorithm to solve it and we also improve the genetic algorithm. Ultimately, the intelligent scheduling subsystem of the production planning is accomplished.Firstly, the genetic algorithm for the flow shop problem is studied, including the genetic coding, fitness function, the parameters of the algorithm, the operators of selection, crossover and mutation, and the terminate condition. To illustrate it, we work out a concrete example of the flow shop problems, using the genetic algorithm. The simulation testing indicates that it has many advantages such as easy coding, simple operating and good optimal performance.Secondly, the hybrid flow shop problem is studied, and an improved genetic algorithm, multi-pairs of chromosomes genetic algorithm is proposed. The algorithm includes a new coding method and a novel crossover operator. Both the simple genetic algorithm and multi-pairs of chromosomes genetic algorithm separately solve a concrete hybrid flowshop problem. The comparison of the simulation results shows that the multi-pairs of chromosomes genetic algorithm is better than the simple genetic algorithm.Thirdly, according to the attributes of the clothing enterprise, its planning problem is simplified to a problem similar to the flow shop problem, the mathematic model is presented, and the concrete genetic algorithm for solving it is studied. The calculation outcome demonstrates that the model and the method are effective.Next, among the ERP system for clothing industry, the intelligent scheduling subsystem of the production planning is implemented. Its function mainly consists of the automatic production planning, the automatic calculation of the materials and so on. Based on the orders and the productivity, the subsystem arranges all the tasks and plays a great role in the decision of the management.Finally, a summary of the main work of this paper is made and the next study of the model, the algorithm and the interface between the ERP for clothing industry and other subsystems like CAD is expected.更多还原