An integrated functional representation of concurrent engineering

A time-based competitive strategy is a driving force for enterprises that are striving to be competitive on a global basis. One of the methods to help in the time compression of the product development life cycle is concurrent engineering. Concurrent engineering is not only an approach that is used by the engineering function of the organization, but it is also one that crosses organizational boundaries. In this article an integrated functional model of the concurrent engineering process and its relationship with the product development life cycle is presented. A functional comparison to the traditional engineering approach for product development is also provided.     

并行工程产品开发过程定量化建模与计划制订

由于现有的一些并行工程定量化模型中修改设计微循环的细节特征描述得还不够充分 ,本文提出了一种新的基于产品 -工艺设计活动对网络的定量化模型来描述并行工程产品开发过程 ,并且给出了产品 -工艺设计活动对平均持续时间与产品或工艺设计活动资源占用率的计算方法 .在此基础上 ,将并行工程产品开发过程的计划制订问题建模成一个有资源约束的项目调度问题 .与其它此类问题不同的是 ,在本文中 ,分配给产品开发项目的各类资源的数量不是事先给定的 ,而是与最优产品开发计划一起获得的 ,所以这是一个资源分配与计划制订的集成优化问题 .本文提出了一种新的基于分枝定界的算法来解决此问题并引入一个启发式规则来提高算法的搜索效率     

An algorithm to dynamically adjust the number of Kanbans in stochastic processing times and variable demand environment

In a real-life environment, the just-in-time JIT system is subjected to various types of uncertainties such as stochastic processing times and variable demand. Since, JIT was only meant to operate in a deterministic environment, its performance is seriously affected by variations in processing times and demand. In this paper, a newly developed Kanban system is presented which uses an algorithm to dynamically and systematically manipulate the number of Kanbans in order to offset the blocking and starvation caused by the said uncertainties during a production cycle. The new system is termed a flexible Kanban system FKS . The steps of the algorithm are detailed and the effectiveness of FKS is demonstrated using an example model. For the example model, the solution procedure, results and a discussion are presented.     

Transforming an MRP plan into a short-term schedule

A new approach for transforming MRP orders, planned periodically, e.g. on a weekly base, into a detailed sequence of jobs is presented. In this model for a single machine environment, the jobs are partitioned into families and a family specific set-up time is required at the start of each period and of each batch, where a batch is a maximal set of jobs in the same family, that are processed consecutively. An integer program is formulated for both the problem of minimizing the number of overloaded periods and the problem of minimizing the total overtime. These programs generate benchmark results for the heuristic approach. A heuristic model is developed that constructs a schedule in which overloaded periods are relieved and set-up time is saved. In this approach, the job sequence is constructed by repeatedly solving a knapsack problem. The weights used in this knapsack problem relate to the preferred priorities of the jobs not yet scheduled and determine the quality of the final sequence. The different features of the heuristic model are compared using a large set of test problems. The results show that the quality of the final sequence depends on an appropriate choice for the weights.     

Centralized nurse scheduling to simultaneously improve schedule cost and nurse satisfaction

Nursing managers are faced with rising turnover and shortages of qualified nursing staff. At the same time they are under increased pressure to simultaneously increase patient care and satisfaction while reducing costs. In this study, we examine the impact of centralizing scheduling decisions across departments in a hospital. By pooling nurses from multiple units and scheduling them in one model, improved costs and reduced overtime result. Reduced overtime improves schedules for nurses. Improved satisfaction levels can positively impact turnover rates among nurses. Our results show that by using a centralized model, nursing managers in hospitals can improve the desirability of nurse schedules by approximately 34% and reduce overtime by approximately 80% while simultaneously reducing costs by just under 11%.     

产品并行开发过程中的风险分析

风险是产品开发的固有特性,本文分析产品并行开发中的风险问题。首先基于上游演化度和下游灵敏度分别建立产品并行开发中的时间函数和成本函数,并结合相关的风险分析,建立产品并行开发的时间风险和成本风险函数,并且基于这两种风险的可转化性,追求其加权和最小,保证产品开发在时间和成本两方面达到预期目标。     

Determination of the optimal number of kanbans and kanban allocation in a FMS: A simulationbased study

A simulation-based study is presented to compare the performance of a set of kanban allocation rules in a flexible manufacturing system (FMS) operating in a pull mode. This paper also presents a simulation-based direct search approach to determine the optimal number of kanbans and their simultaneous allocation to different part types. This paper presents a realistic cost function that includes a penalty for failing to meet the demand on time, and a penalty for rescheduling delivery of materials from vendors. The model captures limited flexibility in the form of operational flexibility in a FMS. Future extensions to this study are discussed.     

The pull evolution: from Kanban to customised token-based systems

This work formalises a new approach for material flow control, i.e. the so-called customised token-based systems (CTBS). This approach represents a natural evolution from classical pull systems, such as Kanban. The article first describes the flow of materials (orders or tasks) from a management point of view, and discusses exact and approximate approaches to tackle the problem. The origin of Customised Production Control systems is explained, as well as the main advantages: high degree of generalisation, enhancement of the space of solutions and adaptation to the manufacturing environment. Next, we propose a method based on cross-entropy for the optimisation of CTBS. The computational experience carried out in a number of scenarios shows the CTBS to be profitable. Furthermore, the resulting control mechanisms seem to be relatively easy to implement in real-manufacturing environments.     

Freezing the master production schedule in multilevel material requirements planning systems under deterministic demand

This paper extends the studies by Sridharan, Berry, and Udayabhanu from single-level MPS systems to multilevel material requirements planning (MRP) systems, and examines the impact of product structure, lot-sizing rules and cost parameters upon the selection of MPS freezing parameters under deterministic demand. A model is built to simulate the master production scheduling and material requirements planning operations in a make-to-order environment. The results show that all the MPS freezing parameters studied have a significant impact upon total inventory costs and schedule instability in multilevel MRP systems. First, the order-based freezing method is preferable to the period-based method. Secondly, the study finds that increasing the freezing proportion reduces both total inventory costs and schedule instability. This finding contradicts the finding by Sridharan et al. in single-level systems. Thirdly, the study finds that a higher replanning periodicity results in both lower total inventory cost and lower schedule instability. The study also indicates that the product structure and lot-sizing rules do not significantly influence the selection of MPS freezing parameters in a practical sense under most situations. However, the cost parameter seems to significantly influence the selection of replanning periodicity.     

Analytic evaluation of the expectation and variance of different performance measures of a schedule on a single machine under processing time variability

In this paper, we present closed-form expressions, wherever possible, or devise algorithms otherwise, to determine the expectation and variance of a given schedule on a single machine. We consider a variety of completion time and due date-based objectives. The randomness in the scheduling process is due to variable processing times with known means and variances of jobs and, in some cases, a known underlying processing time distribution. The results that we present in this paper can enable evaluation of a schedule in terms of both the expectation and variance of a performance measure considered, and thereby, aid in obtaining a stable schedule. Additionally, the expressions and algorithms that are presented, can be incorporated in existing scheduling algorithms in order to determine expectation-variance efficient schedules.     

A dynamic programming approach of finding an optimal broadcast schedule in minimizing total flow time

We study the problem of (off-line) broadcast scheduling in minimizing total flow time and propose a dynamic programming approach to compute an optimal broadcast schedule. Suppose the broadcast server has k pages and the last page request arrives at time n. The optimal schedule can be computed in O(k³(n+k)^k−1) time for the case that the server has a single broadcast channel. For m channels case, i.e., the server can broadcast m different pages at a time where m < k, the optimal schedule can be computed in O(n^k−m) time when k and m are constants. Note that this broadcast scheduling problem is NP-hard when k is a variable and will take O(n^k−m+1) time when k is fixed and m ≥ 1 with the straightforward implementation of the dynamic programming approach. The preliminary version of this paper appeared in Proceedings of the 11th Annual International Computing and Combinatorics Conference as “Off-line Algorithms for Minimizing the Total Flow Time in Broadcast Scheduling”.     

An approach to learning from both good and poor factory performance in a Kanban-based just-in-time production system

In a JIT manufacturing environment it may be desirable to learn from an archived history of data that contains information that reflects less than optimal factory performance. The purpose of this paper is to use rule induction to predict JIT factory performance from past data that reflects both poor (saturated or starved) and good (efficient) factory performance. Inductive learning techniques have previously been applied to JIT production systems (Markham et al. , Computers and Industrial Engineering, 34 , 717-726, 1998; Markham et al. , International Journal of Manufacturing Technology Management, 11 (4), 239-246, 2000), but these techniques were only applied to data sets that reflected a well-performing factory. This paper presents an approach based on inductive learning in a JIT manufacturing environment that (1) accurately classifies and predicts factory performance based on shop factors, and (2) identifies the important relationships between the shop factors that determine factory performance. An example application is presented in which the classification and regression tree (CART) technique is used to predict saturated, starved or efficient factory performance based on dynamic shop floor data. This means that the relationship between the variables that cause poor factory performance can be discovered and measures to assure efficient performance can then be taken.     

An evaluation of flow-shop scheduling algorithms for makespan reduction in a stochastic environment

This paper addresses the problem of makespan reduction and improvement in related performance measures in the stochastic flow shop. The experimental design addresses the issues of the problem size in terms of the number of jobs and machines, the bottleneck location within the production facility, and the processing time distribution and sensitivity to variance. In other words, many of the assumptions that are typically made in the published literature are violated in favour of a more realistic production basis. Experiments are performed via simulation to examine the performance of several well known flow-shop scheduling algorithms and one new algorithm in this challenging environment. The authors conclude that distributional effects and bottleneck considerations can play a role in the performance of the various algorithms considered. This paper further indicates that the problem size also tends to drive the effectiveness of the scheduling strategies examined, and presents information regarding interesting interaction effects between the problem size and the other elements of experimental concern.     

Two approaches to solving the multi-depot vehicle routing problem with time windows in a time-based logistics environment

This paper presents a two-phase heuristic method that can be used to efficiently solve the intractable multi-depot vehicle routing problem with time windows. The waiting time that was ignored by previous researchers is considered in this study. The necessity of this consideration is verified through an initial experiment. The results indicate that the waiting time has a significant impact on the total distribution time and the number of vehicles used when solving test problems with narrow time windows. In addition, to fairly evaluate the performance of the proposed heuristic method, a meta-heuristic method, which extends the unified tabu search of Cordeau et al., is proposed. The results of a second experiment reveal that the proposed heuristic method can obtain a better solution in the case of narrow time windows and a low capacity ratio, while the proposed meta-heuristic method outperforms the proposed heuristic method, provided that wide time windows and a high capacity ratio are assumed. Finally, a well-known logistics company in Taiwan is used to demonstrate the method, and a comparison is made, which shows that the proposed heuristic method is superior to the current method adopted by the case company.     

Analysis of Local Decision Rules in a Dual-Kanban Flow Shop

A recent article reported the results of a study on the effects of two kanban policy variables—the length of withdrawal cycle and the type of priority rule—on average customer wait time and total inventory. This study extends that work by adding two kanban policy variables and two performance criteria. It reports the results of simulation experiments that were conducted to determine how four policy variables—withdrawal cycle, priority rule, status of waiting withdrawal kanbans, and number of kanbans influence four performance criteria—average customer wait-time, total inventory, and average number of full containers in the input and output stock points of stations. It was found that the information about waiting withdrawal kanbans in sequencing decisions results in the simultaneous improvement in two conflicting objectives—customer wait time and total inventory. Also, the effects of including the information regarding the status of waiting withdrawal kanbans on system performance are larger than the effects associated with the type of priority rule. The results provide insights into determining the level of each policy variable while fully considering the possible interactions among the variables and the levels of other policy variables to improve system performance. These insights allow for setting the levels of policy variables to make the improvement process smooth.     

随机环境中的生产作业计划问题

生产系统中通常会涉及各种不确定因素 ,如不确定的顾客定单、不确定的生产作业时间等 .在当今时间竞争非常激烈的时代中 ,生产型企业如何把握生产系统中的这些不定因素变得尤为关键 .本文研究在不确定的作业时间、工序间延迟时间等情况下的生产作业计划问题 ,利用 scenario模型把这类随机生产计划问题归纳为一个多阶段随机决策问题 .进而 ,采用Lagrangian松弛和 scenario分解的方法求解这样一个大型的决策问题 .最后 ,就一个实例建立模型、进行计算和分析 ,以说明本文提出的随机生产计划方法的特点和有效性     

Card-based production control: a review of the control mechanisms underpinning Kanban,ConWIP, POLCA and COBACABANA systems

Since the emergence of Kanban, there has been much research into card-based control systems. This has included attempts to improve Kanban and/or develop alternative systems, particularly ConWIP (i.e. Constant Work-In-Process), POLCA (i.e. Paired-cell Overlapping Loops of Cards with Authorisation) and COBACABANA (i.e. Control of Balance by Card-Based Navigation). Yet, to date, no unifying review of the mechanisms underpinning these systems has been presented. As a consequence, managers are not provided with sufficient support for choosing an appropriate system for their shop; and researchers lack a clear picture of how the mechanisms compare, leading to several misconceptions. This paper reviews the control mechanisms underpinning the Kanban, ConWIP, POLCA and COBACABANA systems. By comparing the ‘control mechanism’ (i.e. the loop structure and card properties) and ‘contextual factors’ (i.e. routing variability, processing time variability, and whether stations are decoupled by inventory or the flow of jobs is controlled), we provide managers with guidance on which system to choose. For research, we show for example that most criticisms put forward against Kanban systems, e.g. to justify the development of ConWIP, POLCA or COBACABANA, only apply to work-in-process Kanban systems and not to production Kanban systems. Future research directions for each control system are outlined.     

Applying the Partitioned Multiobjective Risk Method (PMRM) to Portfolio Selection

The analysis of risk-return tradeoffs and their practical applications to portfolio analysis paved the way for Modern Portfolio Theory (MPT), which won Harry Markowitz a 1992 Nobel Prize in Economics. A typical approach in measuring a portfolio's expected return is based on the historical returns of the assets included in a portfolio. On the other hand, portfolio risk is usually measured using volatility, which is derived from the historical variance-covariance relationships among the portfolio assets. This article focuses on assessing portfolio risk, with emphasis on extreme risks. To date, volatility is a major measure of risk owing to its simplicity and validity for relatively small asset price fluctuations. Volatility is a justified measure for stable market performance, but it is weak in addressing portfolio risk under aberrant market fluctuations. Extreme market crashes such as that on October 19, 1987 ("Black Monday") and catastrophic events such as the terrorist attack of September 11, 2001 that led to a four-day suspension of trading on the New York Stock Exchange (NYSE) are a few examples where measuring risk via volatility can lead to inaccurate predictions. Thus, there is a need for a more robust metric of risk. By invoking the principles of the extreme-risk-analysis method through the partitioned multiobjective risk method (PMRM), this article contributes to the modeling of extreme risks in portfolio performance. A measure of an extreme portfolio risk, denoted by f(4), is defined as the conditional expectation for a lower-tail region of the distribution of the possible portfolio returns. This article presents a multiobjective problem formulation consisting of optimizing expected return and f(4), whose solution is determined using Evolver-a software that implements a genetic algorithm. Under business-as-usual market scenarios, the results of the proposed PMRM portfolio selection model are found to be compatible with those of the volatility-based model. However, under extremely unfavorable market conditions, results indicate that f(4) can be a more valid measure of risk than volatility.     

Integrating Kanban principles in a pharmaceutical campaign production system

Centralised planned campaign production is the predominant production system in process industries. In this study, we investigate whether a decentralised Kanban control system, which has proven to offer advantages in other industries, can be successfully integrated in a campaign production environment. The used research methodology combines model-based and case-study-based elements. Using the example of a pharmaceutical internal five-stage supply chain, we conceive a Kanban concept that integrates campaign formation, develop a discrete-event simulation model and conduct a range of explorative simulation experiments. We find that a Kanban campaign production system is not merely feasible but would also be favourable: throughput times can be reduced without increasing customer lead times. Sensitivity analysis shows that the system’s performance is relatively robust to changes in Kanban key configuration parameters such as number of Kanban cards or campaign size. We conclude by discussing our findings and formulating three propositions that might stimulate future research.     

大规模集成电路预烧作业中分批排序问题的数学模型

分批排序(Batch Scheduling)是在半导体生产过程的最后阶段提炼出来的一类重要的排序问题。单机分批排序问题就是n个工件在一台机器上加工,要将工件分批,每批最多可以同时加工B个工件,每批的加工时间等于此批工件中的最大的加工时间。Skutella^[8]1998年把平行机排序的P｜｜∑ω_jC_j和R｜｜∑ω_jC_j表述成二次的0-1整数规划,得到一些令人满意的结果;国内罗守成等^[9]、张倩^[10]给出了单机排序问题1｜｜∑ω_jC_j的数学规划表示,对于用数学规划来研究排序问题是一个很有意义的进展。本文首先介绍总完工时间和最小的带权单机分批排序问题1｜B｜∑ω_jC_j,然后将1｜B｜∑ω_jC_j表示成数学规划的形式,并且用数学规划中的对偶理论证明了SPT序是其特殊情况1｜B=1｜∑C_j的最优解。