需求不确定环境下可调整产能的混流装配线平衡研究

发布时间：2018-05-19 17:30

本文选题：混流装配线 + 装配线平衡　；参考：《西安交通大学》2017年博士论文

【摘要】：混流装配线(以下简称混装线)能同时装配多种型号的产品,因此被广泛的应用于汽车、家电和消费电子等行业,以满足多样化的顾客需求。装配线平衡是产线设计的核心决策,它决定各工作台的任务分配,直接影响装配线的效率和成本。此问题有超过50年的研究历史,现有研究中几乎都是按固定产能,或者说按预计的生产需求来设计装配线的。然而在实际中,为应对瞬息万变的市场需求,企业经常需要调整生产计划,改变装配线的生产任务。当实际的生产需求偏离设计值时,装配线可能无法达到设计的生产效率,偏离较大时,可能还须进行相应的调整措施(例如加班,临时工等)才能满足生产需求。调整发生的频率和产生的额外成本受到需求波动的程度和初始平衡方案的影响,但是这些因素在现有平衡研究中极少考虑,从而导致生产成本的低估。为弥补这种缺陷,本文研究了不确定环境下的混装线平衡问题,混装线可以采取一些产能调整措施以应对需求的变化。通过在平衡阶段就考虑后续产能调整的便利性和成本,可以降低在产线在不确定需求环境中长期运行的平均成本。本文讨论了三种产能调整措施,针对不同的措施分别研究了相应的产线平衡问题,主要工作和创新点归纳如下:1.研究了使用备用多能工调整产能的混装线平衡问题。使用多能工是应对工作台负荷过载最常用的方法之一,在企业实践中也很常见。许多混装线排序问题的研究已讨论了这种方法,但还没有研究讨论它对混装线平衡的影响。备用多能工扮演一个产能缓冲池的作用,他们平时只少量或不参与装配作业,当有工作台发生负荷过载时他们才去协助装配。与普通工人不同的是,多能工能帮助多个工作台,工资更贵。决策者需要决定装配线使用的普通工人和备用多能工的数量、以及各工作台的任务分配,以满足任意情境的生产需求。本文以最小化总人工成本为目标建立了该问题的数学模型,根据问题特点,提出了一种递归算法计算给定普通工人数量时过载工作量的下界。在此基础上,给出了全局成本下界的估计方法。本文提出了一种启发式算法快速寻找近似最优解,并以此为上界,设计了分支定界与记忆(BBR)算法精确求解该问题。本文基于SALBP-1标准算例库(http://alb.mansci.de/)随机生成了500个算例以验证算法有效性,计算实验表明,算法能在60秒内获得并验证其中406个问题的最优解,对未能验证最优性的94个算例,获得的解距离下界的平均偏差为5.17%。数值实验还比较了固定产能和允许产能调整的成本,结果表明使用备用多能工能够实现平均幅度为5%左右的成本改善,但多能工难以随时雇佣的特点限制了这种方法的适用范围,只有小部分具备特定特征的算例会采用这种产能调整方法。2.研究了通过加班调整产能的混装线平衡问题。加班是企业实践中最常用的临时改变产能的方法,但还没有研究考虑允许加班对混装线平衡方案的影响。加班获得的额外劳动时间可临时提高混装线的生产能力,但加班需要支付更高的工资,且最大加班时间受法律限制。决策者需要决定使用的工作台数量、各工作台的任务分配和各情境下的加班时间,以满足每个情境的需求,并最小化平均每天需支付的总工资。本文建立了该问题的数学模型,分析了问题的若干性质,提出了计算全局成本下界的迭代算法。之后,利用下界计算方法,提出了启发式算法和BBR算法来求解本问题。对500个算例的计算实验表明,BBR算法能在60秒内精确求解其中408个,对未能获得最优的92个算例,获得的解离下界的平均偏差为2.67%。数值实验还比较了固定产能和允许产能调整的成本,结果表明,绝大多数算例都能够通过加班实现较明显的成本改善,平均改善幅度随着需求波动程度增大而增大,但即便需求变化不超过20%,成本改善的幅度也超过9%。3.研究了通过增减工人数量调节产能的混装线平衡问题。Simaria等(2009)[3]研究了U型线常见的一种产能调节方法:不移动设备位置,通过改变工人数量和每个人的任务分配来应对需求的变化。本文讨论了这种方法在直线型混装线中应用:保持任务顺序不变,通过重新划分工作台边界的方式改变工作台数量。决策者需要决定任务顺序和各情境的工作台划分,以减少各情境使用的工作台数量的平均值。本文对此问题建立了数学模型。分析了给定序列下最优工作台划分的性质,给出了下界的计算方法,然后提出了基于位置交换的局部优化方法,并将之嵌入到单程启发式算法中,最后提出了BBR精确算法。对500个算例的计算实验表明,BBR算法获得了308个问题的精确最优解,192个未验证算例与成本下界的平均偏差为4.05%。数值实验还比较了固定产能和允许产能调整的成本,结果表明,绝大多数算例都能够通过通过增减工作台数量减少所需人工成本,成本改善幅度随着需求波动程度增大而增大。当需求波动超过50%时,这种方法能实现比前两种方法更明显的成本改善;即便需求变化不超过20%,成本改善的幅度也超过7%。
[Abstract]:Mixed assembly line (hereinafter referred to as mixed line) can assemble a variety of types of products at the same time, so it is widely used in automotive, household electrical appliances and consumer electronics industries to meet the diverse needs of customers. Assembly line balance is the core decision of line production design. It determines the assignment of tasks in each workbench and directly affects the efficiency and cost of the assembly line. This problem has more than 50 years of research history. In the current research, assembly lines are designed in terms of fixed capacity or expected production demand. In practice, in order to cope with the fast changing market demand, enterprises often need to adjust production plans and change the production tasks of assembly lines. When actual production demand deviates from the design. At the time of value, the assembly line may not be able to achieve the production efficiency of the design. When the deviation is large, the corresponding adjustment measures (such as overtime, temporary workers, etc.) will be required to meet the production needs. The frequency and the additional cost of the adjustment are affected by the degree of demand fluctuation and the initial balance scheme, but these factors are in the current balance. In order to make up for this defect, this paper studies the problem of the mixing line balance in the uncertain environment. The mixed line can take some capacity adjustment measures to cope with the change of demand. The average cost of long-term operation in the uncertain demand environment. Three kinds of capacity adjustment measures are discussed in this paper. According to different measures, the corresponding balance problem of line production is studied. The main work and innovation points are summarized as follows: 1. the problem of mixing line balance using standby multi energy adjusts the productivity is studied. The use of multi energy is to deal with the workbench. One of the most commonly used methods of load overload is also common in enterprise practice. Many research on mixed line sorting problems have been discussed, but there is no study and discussion of its impact on the balance of the mixed line. Standby multiplayer plays a role of a capacity buffer pool, they usually only have little or no participation in assembly operations, when there is a worktable. Unlike ordinary workers, they can help multiple workstations to be more expensive. Decision makers need to determine the number of ordinary workers and spare multiple workers used on the assembly line, and the assignment of tasks in each workbench to meet the production requirements of the task situation. This article minimizes the total labor force. On the basis of the characteristics of the problem, a mathematical model of the problem is set up. Based on the characteristics of the problem, a recursive algorithm is proposed to calculate the lower bound of the overload workload for a given number of ordinary workers. On this basis, the estimation method of the global cost lower bound is given. In this paper, the branch and bound and memory (BBR) algorithm is considered to solve the problem accurately. Based on the SALBP-1 standard example library (http://alb.mansci.de/), 500 examples are randomly generated to verify the effectiveness of the algorithm. The calculation experiments show that the algorithm can obtain and verify the optimal solution of 406 of the 406 problems within 60 seconds, and obtain 94 examples of the failure to verify the optimality. The average deviation of the lower bounds of the solution distance is the 5.17%. numerical experiment, which also compares the cost of the fixed capacity and the allowable capacity adjustment. The result shows that the cost improvement of the average amplitude is about 5% by using the spare multi energy worker, but the characteristics of the multi energy workers' difficult to employ at any time limit the scope of the application of this method, only a small part has specific special characteristics. This capacity adjustment method,.2., is used to study the problem of mixing line balance by overtime adjustment. Overtime is the most commonly used method of temporary change in capacity in enterprise practice, but there is no study to consider the effect of overtime on the mixed line balance scheme. Extra working hours obtained by overtime can temporarily increase the mix line. The production capacity, but overtime pay higher wages, and the maximum overtime time is limited by the law. Decision makers need to determine the number of workstations used, the assignment of the workstations and the overtime hours in each situation to meet the needs of each situation and minimize the average daily wages to be paid. This paper establishes the problem. In the mathematical model, we analyze some properties of the problem and propose an iterative algorithm for calculating the lower bound of global cost. After that, the heuristic algorithm and BBR algorithm are proposed to solve the problem by using the lower bound calculation method. The calculation experiments of 500 examples show that the BBR algorithm can solve 408 of them in 60 seconds, and for the failure to obtain the optimal calculation. For example, the average deviation of the dissociation lower bounds obtained by the 2.67%. numerical experiment also compares the cost of the fixed capacity and the allowable capacity adjustment. The results show that most of the examples can achieve a more obvious cost improvement through overtime, and the average improvement increases with the increase of demand volatility, but even if the demand change does not exceed 20%, the cost can be changed. The extent of the improvement is also more than 9%.3.'s study of the mixed line balance problem (.Simaria) that regulates the productivity by increasing the number of workers. (2009) [3] has studied a common method of capacity adjustment for the U line: the position of the non mobile equipment, the change in the demand by changing the number of workers and the assignment of each person. This method is discussed in this paper. The application of line type mixing line: keeping the order of the task constant and changing the number of worktable by redividing the boundary of the worktable. The decision maker needs to determine the task order and the worktable division of each situation so as to reduce the average number of the worktable used in each situation. This paper establishes a mathematical model for this question and analyzes the given sequence. The nature of the optimal table division gives the calculation method of the lower bound, then puts forward the local optimization method based on the position exchange, and embeds it into the single path heuristic algorithm, and finally puts forward the exact BBR algorithm. The calculation experiments of the 500 examples show that the BBR algorithm obtains the exact optimal solution of the 308 problems and 192 unverified examples. The average deviation from the lower bound of the cost is the 4.05%. numerical experiment that also compares the cost of fixed capacity and allowable capacity adjustment. The results show that most of the examples can reduce the labor cost by reducing the number of worktables, and the cost improvement increases with the increase of demand volatility. The method can achieve more obvious cost improvement than the first two methods. Even if the demand change is not more than 20%, the cost improvement will exceed 7%.
【学位授予单位】：西安交通大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TB497

【相似文献】