不确定环境下生产计划和调度的研究

发布时间：2018-03-13 14:52

本文选题：生产计划　切入点：生产调度　出处：《浙江大学》2016年博士论文　论文类型：学位论文

【摘要】：生产计划和调度不仅是企业生产和管理的重要环节之一,而且是生产决策的有效依据,计划和调度方案的优劣决定了企业的经济效益。生产计划和调度横向覆盖的范围广,纵向包括的层次多,涉及很多不确定因素,例如：产品需求、产品销售价格、生产成本、公用工程、处理时间、生产能力、交货时间等不确定因素。以上不确定性因素对生产计划和调度有重要的影响,可能使优化结果不是最优解甚至是非可行解,从而严重影响企业的生产效率和经济效益。本论文研究不确定环境下的生产计划和调度问题,考虑的不确定因素包括：产品需求、产品销售价格、成本、公用工程等。根据不确定因素的时间尺度,将不确定因素划分到计划层或调度层中进行区别处理。与生产计划相比,生产调度涉及的约束较多,针对生产调度问题的建模方法也更多,所以不确定环境下的生产调度问题是本文的重点研究内容之一。另外,公用工程不确定性与生产过程联系紧密,对调度优化结果的可行性和最优性影响较大,但又经常被忽视。因此,本论文在调度层主要围绕公用工程的不确定问题展开研究,分别从建模方法和生产过程类型两个方面进行了深入的研究和分析。为了降低公用工程不确定性对生产和经济利益的影响,分别提出变周期调度策略、公用工程与生产率之间模糊关系、混合双层连续模型、计划调度双层集成模型等方法。本文的主要研究工作和创新点如下：1)在计划模型中考虑多种不确定因素,例如：产品需求、产品销售价格和成本的不确定性,采用离散时间建模方法,建立混合整数线性规划(Mixed Integer Linear Programming, MILP)不确定计划模型,并利用机会约束规划理论描述以上3种不确定因素。除了基本约束之外,在不确定模型中还考虑了生产率波动和库存水平超限约束。分析了置信因子的不同取值对优化结果的影响及原因。验证了所建模型可以处理带有多种不确定参数的计划问题,使含有不确定参数的约束在大于某概率水平下成立,优化结果为企业在不同战略目标下的生产决策提供依据。2)针对间歇生产过程和连续生产过程,采用离散时间建模方法,分别建立MILP不确定调度模型,用于降低公用工程扰动对生产过程造成的影响。在间歇生产调度模型中,引入四维0-1变量,使得生产任务可以跨越多个调度周期,有效地解决了公用工程统计量不准确的问题。在连续生产调度模型中,提出变周期调度策略,解决了调度周期与公用工程扰动的持续时间不匹配以及计算复杂度问题。算例验证了间歇过程调度模型在公用工程无扰动情况下,可以准确地统计公用工程消耗量；在公用工程带有扰动的情况下,模型的优化结果可以对公用工程扰动做出响应,降低了公用工程扰动对优化结果的影响。连续过程算例验证了当公用工程扰动小于调度周期时,变周期调度模型能够对扰动做出响应,并获得更好的优化结果。3)针对间歇生产过程的不确定调度问题,基于特定单元事件点的连续时间建模方法,建立混合整数非线性规划(Mixed Integer Nonlinear Programming, MINLP)不确定调度模型。采用模糊理论描述公用工程的不确定性,并建立公用工程和生产率之间的模糊线性关系,来间接描述处理时间的不确定性。最后利用加权平均法将不确定模糊模型转化成确定性模型再求解。通过算例验证了模型的有效性,并详细地分析了一些重要的参数对优化结果的影响及其取值方法。4)特定单元事件点连续时间模型(Unit-specific Event-based Continuous Time Model, UEBCTM)存在约束过于严格和计算不准确问题。为了解决UEBCTM模型的缺陷,提出双层模型结构并建立基于特定单元事件点和时间槽的混合模型。采用UEBCTM建模方法,建立上层MILP模型；采用基于时间槽的连续时间建模方法,建立下层的非线性模型。此外,在下层模型中引入模糊理论来描述公用工程供给量的不确定性。双层混合模型的求解通过上、下层模型的迭代获得。算例验证了双层混合模型不但具有UEBCTM模型的优点,而且解决了UEBCTM模型的缺陷,并能够在公用工程不确定的情况下为生产决策提供依据。5)考虑生产计划和调度中的多种不确定因素,建立多周期的计划和调度双层集成模型。计划层采用离散时间建模方法并引入机会约束随机规划来描述需求不确定性,建立离散时间线性不确定计划模型,同时考虑了生产率波动和参考库存约束。调度层解决带有公用工程不确定的生产调度问题,采用UEBCTM建模方法并引入模糊理论建立关于多用途间歇过程的连续时间不确定调度模型。利用滚动优化策略对计划和调度双层模型进行迭代求解。仿真结果验证了所提方法的有效性和优越性。机会约束方法可以结合不确定信息优化产品的产量,为决策者提供更准确的参考依据,有利于决策的合理性和有效性。使用模糊理论表示公用工程的不确定性更真实地描述了公用工程的供给量,增加了调度结果的可行性,降低了不确定参数对优化结果的影响,同时提高了设备和公用工程的利用率。
[Abstract]:Production planning and scheduling is not only an important part of enterprise production and management, and is effective on the basis of production decision, planning and scheduling scheme determines the economic benefits of the enterprise. The scope of production planning and scheduling for horizontal coverage is wide, including multiple levels, involving many uncertain factors, such as product demand product sales prices, production costs, utilities, processing time, production capacity, delivery time and other uncertain factors. These uncertainty factors have an important effect on the production planning and scheduling, may make the optimization results than the optimal solution or even infeasible solutions, thus seriously affect the production efficiency and economic benefits. This paper the uncertainty of production planning and scheduling problem under uncertainty, including consideration: product demand, product sales price, the cost of utilities, etc. according to the uncertain factors The time scale, the uncertain factors into plan layer or layer scheduling distinction. Compared with the production plan, production scheduling involves many constraints, the modeling method for production scheduling is more uncertain, so the production scheduling problem under the environment is one of the key research content of this paper. In addition, the utility is not the uncertainty and the production process is closely linked, has great influence on the optimization results of feasibility and optimality, but often overlooked. Therefore, this paper mainly focus on scheduling layer utilities uncertain issues research, from two aspects of modeling method and process type is studied and analyzed in order to. To reduce the uncertainty of utilities impact on production and economic benefits, propose variable cycle scheduling strategy, the fuzzy relation between utilities and productivity, mixed double continuous Planning and scheduling model, double integration model and method. The main research work and innovations are as follows: 1) in the planning model considering various uncertain factors, such as product demand, product sales price and cost uncertainty, the discrete time modeling method, establish a mixed integer linear programming (Mixed Integer Linear Programming. MILP) uncertain planning model, and use the chance constrained programming theory described above 3 kinds of uncertain factors. In addition to the basic constraints, in the uncertain model also considers the productivity fluctuation and inventory levels exceeding limit constraints. Analysis of the different values of the confidence factor to influence the optimization results and the reasons. The model can be verified treatment plan with a variety of uncertain parameters, the uncertain parameters in the establishment of a level greater than the probability, the results of the optimization of enterprises in different strategic goals The production decision.2) provide the basis for batch process and continuous process, the discrete time modeling method, MILP uncertain scheduling model are established respectively, to reduce the disturbance impact on the utility of the production process. In the batch production scheduling model, introducing the four-dimensional 0-1 variables, making the production task spanning multiple scheduling cycle, effectively solves the problem of inaccurate statistics of public works. In the continuous production scheduling model, put forward the periodic scheduling strategy, solve the duration of scheduling period and utility disturbances do not match and the computational complexity problem. Examples show the batch process scheduling model of non disturbance in utilities, can accurate statistical utility consumption; with disturbances on the utility model, the optimization results can respond to public engineering disturbance reduction The utility of low disturbance influence on the optimization results. The continuous process is verified when less than the scheduling period utility disturbance, variable cycle scheduling model can respond to disturbance, and obtain better optimization results.3) for uncertain scheduling of batch production process, continuous time modeling method based on the specific unit event, a mixed integer nonlinear programming (Mixed Integer Nonlinear Programming, MINLP) uncertain scheduling model. Using fuzzy theory to describe the utility of uncertainty, fuzzy linear relationship and the establishment of utilities and productivity between the indirect description of processing time uncertainty. The uncertainty of the fuzzy model is transformed into a deterministic model and then solved by the weighted average method. Through an example to verify the validity of the model, and a detailed analysis of some important parameters on the optimization results. Sound and value method.4) specific unit event continuous time model (Unit-specific Event-based Continuous Time Model, UEBCTM) constraint is too strict and not accurate calculation. In order to solve the defects of the UEBCTM model, the double model structure and a hybrid model specific unit event and time slot based. Using UEBCTM modeling method, establish the upper MILP model; continuous time modeling method based on time slots, a nonlinear model of the lower. In addition, fuzzy theory is introduced to describe the utility supply uncertainty in the lower model. The bilevel mixed model by iteration, obtain the lower model. Examples show the advantages of hybrid model not only has the UEBCTM the model, and solves the defects of the UEBCTM model, and in the public works under the condition of uncertainty is provided according to the production decision According to.5) consider a variety of production planning and scheduling uncertain factors in the planning and scheduling model of multi period integrated double layer. The plan by using the discrete time modeling method and the introduction of chance constrained programming to describe the demand uncertainty, a discrete time uncertain linear planning model, considering the productivity fluctuations and reference inventory constraints. Scheduling layer to solve scheduling problem with uncertain utilities, using UEBCTM modeling method and fuzzy theory based on continuous time multipurpose batch process uncertain scheduling model is introduced. The rolling optimization strategy for planning and scheduling of two-layer model can be used in iteration. Simulation results verify the validity and superiority of the proposed method.. the chance constraint method can be combined with uncertain information to optimize the output of product, to provide more accurate reference for decision makers, is conducive to the decision Reasonable and effective. The use of fuzzy theory to describe the utility uncertainty more realistically describes the supply of utilities, increase the feasibility of scheduling results, reduce the influence of uncertain parameters on the optimization results, and improve the utilization rate of equipment and utilities.

【学位授予单位】：浙江大学
【学位级别】：博士
【学位授予年份】：2016
【分类号】：TB497

【相似文献】