Evaluation of three production planning procedures for the use of recipe flexibility

Process industries often obtain their raw materials from mining or agricultural industries. These raw materials usually have variations in quality, which often lead to variations in the recipes used for manufacturing a product. Another reason for varying the recipe is to minimize production costs by using the cheapest materials that still lead to a satisfactory quality in the product. A third reason for using recipe flexibility is that it may occur that at the time of production not all materials for the standard recipe are available. In earlier research we showed under what conditions the use of this type of recipe flexibility should be preferred to the use of high materials stock to avoid materials shortages. We also showed that the use of recipe flexibility to account for material shortages can be justified if the material replenishment leadtime is long, the demand uncertainty is high and the required service level is high. In this paper we assume that these conditions are satisfied and we investigate three different production planning procedures that make use of recipe flexibility to cope with the uncertainty in demand and supply. We assume that the customer order leadtime is much smaller than the material replenishment leadtime, and therefore demand uncertainty is high. The optimal procedure optimizes material use over a planning horizon equal to the material replenishment leadtime, taking into account the customers orders and knowledge of the distribution function of future demand. The deterministic procedure also optimizes the material use over the material replenishment leadtime, but it assumes a deterministic demand level for unknown orders. The simplest, myopic procedure optimizes material use over only the accepted customer orders. These three procedures are investigated via an experimental design of computer simulations of an elementary small scale model of the production planning situation. The results show that the optimal procedure outperforms the other two procedures. Furthermore, for a realistic cost structure in feed industry under certain circumstances the use of the optimal procedure may lead to a 4% increase in profit. However, this improvement must be weighted against the cost incurred by the operational use of this complex procedure. Based on these considerations and the numerical results in this paper, we may expect that for some situations in practice the use of the simplest myopic procedure, optimizing material use only over the available customer orders, will be justified from an overall cost point of view.     

Predictive–reactive scheduling on a single resource with uncertain future jobs

We consider a scheduling problem where the firm must compete with other firms to win future jobs. Uncertainty arises as a result of incomplete information about whether the firm will win future jobs at the time the firm must create a predictive (planned) schedule. In the predictive schedule, the firm must determine the amount of planned idle time for uncertain jobs and their positions in the schedule. When the planned idle time does not match the actual requirements, certain schedule disruptions occur. The firm seeks to minimize the sum of expected tardiness cost, schedule disruption cost, and wasted idle time cost. For the special case of a single uncertain job, we provide a simple algorithm for the optimal planned idle time and the best reactive method for schedule disruptions. For the case of multiple uncertain jobs, a heuristic dynamic programming approach is presented.     

Deterministic/stochastic assumptions in job shops

The study investigates the effects which possibly unrealistic assumptions of accurately predicting operation times may have on relative performance of various job shop dispatching rules as compared with using an assumption of not being able to accurately predetermine such times. The experimental design includes factors dealing with the amount of accuracy in the estimated operation times, job dispatching heuristic rules, and shop loading categories. The stochastic operation times represent two different degrees of inaccuracy; one level reflects an estimated ‘normal’ amount of inaccuracy associated with an experienced predictor (shop foreman) while the other level doubles the amount of variance associated with the ‘normal’ predictor's error. These two stochastic levels are compared to a deterministic level where predetermined operation times are absolutely accurate. Five different heuristic rules are evaluated under six different shop loading levels. General conclusions indicate that an assumption of accurately predetermining actual operation times is not likely to weaken the analysis and impact of the research studies which have been performed using such an assumption. However, a specific conclusion indicates that, for at least one shop loading category, researchers should be careful when extending conclusions based on one operation time assumption to situations involving the other assumption.     

Manufacturing delivery performance for supply chain management

This paper considers single-stage make-to-order production systems. We focus on (1) modeling the appropriate expected costs under a variety of modeling assumptions and (2) characterizing the optimal policies. Our approach to solving the problem is to derive the distribution of actual completion times of the process for individual orders and to compare it to the corresponding quoted due dates in order to obtain the expected total costs. We then show the convexity of the objective cost function for determining the decision variable(s), the planned customer order leadtime.     

Perishable inventory management subject to stochastic leadtime

This paper discusses the model for a perishable product with the stochastic procurement leadtime. Zero or 1 unit leadtime is considered and its optimal ordering policy is derived. This is a generalization of the one period horizon model of Nahmias [4]. Further, the chance constraint as for shortage is added and under this additional constraint, again optimal ordering policy is obtained.     

供应提前期和需求不确定下库存系统的随机比较

本文研究供应提前期和需求不确定下包含最小化成本和最大化利润的一致化报童问题中随机比较实现的充分条件,并通过随机比较定量刻画提前期和需求不确定性对库存系统决策和利润的影响。首先引入报童收益函数,给出其优化前后满足随机占优及拉普拉斯变换序的充分条件。进一步,在周期需求服从指数分布的情形下,定量刻画固定提前期与提前期服从几何分布两种情况的随机单调性。最后,将上述结论应用到报童模型中,得到提前期需求对一致化报童问题的随机单调性。     

Simultaneous optimization of capacity and planned lead time in a two-stage production system with different customer due dates

We consider a two-stage make-to-order manufacturing system with random demands, processing times, and distributed customer due dates. The work to each stage is released based on a planned lead time. A general approach to minimize total inventory holding and customer order tardiness cost is presented to find the optimal manufacturing capacities and planned lead times for each manufacturing stage. Expressions are derived for work-in process inventories, finished-goods-inventory and expected backorders under the assumption of a series of M/M/1 queuing systems and exponentially distributed customer required lead times. We prove that the distribution of customer required lead time has no influence on the optimal planned lead times whenever capacity is predefined but it influences the optimal capacity to invest into. For the simultaneous optimization of capacity and planned lead times we present a numerical study that shows that only marginal cost decreases can be gained by setting a planned lead time for the upstream stage and that a considerable cost penalty is incurred if capacity and planned lead time optimization are performed sequentially.     

Bivariate Forecasting with a Variable Leadtime

The assumption of a constant leadtime between a forecast variable and a leading indicator is difficult to support in a number of forecasting contexts, especially in relationships between economic variables. The forecasting method described uses a leading indicator and has the capability to handle varying leadtimes between the forecast variable and the leading indicator. This is achieved by modelling the behaviour of the leadtime as a Markov chain. The method is shown to be effective at detecting and reacting to changes in the leadtime by using simulated series, the only situation where the leadtime is known, and demonstrated on other series.     

具有不可用区间且工件可拒绝下的单机重新排序问题的近似方案

本文考虑了机器具有不可用区间且工件可拒绝下的单机重新排序问题,在该问题中,给定一个工件集需在一台机器上加工,每个工件有自己的加工时间和权重,且对该工件集目标函数为极小化总加权完工时间的排序计划已给定,根据该排序计划中每个工件的完工时间已确定每个工件的承诺交付时间。然而,在工件正式开始加工前,原计划用于加工的某段时间区间因临时用于检修机器而导致机器在该时间区间不再可用,需要对工件重新排序。为了确保在新的重新排序中,工件的延误成本不致太大,决策者可以选择拒绝部分工件,但需支付相应的拒绝费用。任务是确定接受工件集和拒绝工件集,并将接受的工件在考虑机器具有不可用区间的条件下重新排序使得接受工件集的总加权完工时间,总拒绝费用及赋权最大延误之和最小。该问题是NP-困难的,对此给出了伪多项式时间动态规划精确算法,利用稀疏技术设计了完全多项式时间近似方案。     

On a borderline between the NP-hard and polynomial-time solvable cases of the flow shop with job-dependent storage requirements

The paper is concerned with the two-machine flow shop, where each job requires an additional resource (referred to as storage space) from the start of its first operation till the end of its second operation. The storage requirement of a job is determined by the processing time of its first operation. At any point in time, the total consumption of this additional resource cannot exceed a given limit (referred to as the storage capacity). The goal is to minimise the makespan, i.e. to minimise the time needed for the completion of all jobs. This problem is NP-hard in the strong sense. The paper analyses how the parameter - a lower bound on the storage capacity specified in terms of the processing times, affects the computational complexity.

     

Stochastic leadtimes in a one-warehouse,N-retailer inventory system with the warehouse not carrying stock

This study extends upon a multi-echelon inventory model developed by Graves, introducing in the one-warehouse, N-retailer case—as Graves suggested—stochastic leadtimes between the warehouse and the retail sites in place of the original deterministic leadtimes. Effects of stochastic leadtimes on required base stock levels at the retail sites in the case where the warehouse carries no stock (e.g., serves as a cross-dock point) were investigated analytically. Two alternative treatments of stochastic leadtime distributions were considered. Using as a baseline Graves’ computational study under deterministic leadtimes, results of the current study suggest that it may be better to use the deterministic model with an accurately estimated mean leadtime than a stochastic model with a poorly estimated mean leadtime.     

Congestion-based leadtime quotation and pricing for revenue maximization with heterogeneous customers

This paper studies a queuing model where two customer classes compete for a given resource and each customer is dynamically quoted a menu of price and leadtime pairs upon arrival. Customers select their preferred pairs from the menu and the server is obligated to meet the quoted leadtime. Customers have convex–concave delay costs. The firm does not have information on a given customer’s type, so the offered menus must be incentive compatible. A menu quotation policy is given and proven to be asymptotically optimal under traditional large-capacity heavy-traffic scaling.     

On the (R,s, Q) inventory model when demand is modelled as a compound Bernoulli process

In this paper we present an approximation method to compute the reorder point s in an (R, s, Q) inventory model with a service level restriction. Demand is modelled as a compound Bernoulli process, i.e., with a fixed probability there is positive demand during a time unit; otherwise demand is zero. The demand size and the replenishment leadtime are random variables. It is shown that this kind of modelling is especially suitable for intermittent demand. In this paper we will adapt a method presented by Dunsmuir and Snyder such that the undershoot is not neglected. The reason for this is that for compound demand processes the undershoot has a considerable impact on the performance levels, especially when the probability that demand is zero during the leadtime is high, which is the case when demand is lumpy. Furthermore, the adapted method is used to derive an expression for the expected average physical stock. The quality of both the reorder point and the expected average physical stock, calculated with the method presented in this paper, rum out to be excellent, as has been verified by simulation.     

An exploration of models that minimize leadtime through batching of arrived orders

In the last decade interest in work-in-process inventory control has grown. Many papers deal with this topic by considering the manufacturing leadtime as the critical factor that determines the amount of work-in-process. Several authors studied the influence of a batching decision on the average manufacturing leadtime. To this end queueing models with batch arrivals and batch service times were analyzed. One of the underlying assumptions made in the analysis is that the arrival process of the batches can be approximated by a Poisson process for each choice of the batchsize. However, when the interarrival times of individual clients are negative exponentially distributed an Erland distribution may be more appropriate as distribution of the interarrival time of the batches at the production unit. In this paper we consider the single item case. A very tractable analytical approximation for the average leadtime when batches arrive according to an Erlang distribution will be derived. Expressions for the optimal batchsize and the associated minimal leadtime are calculated and compared to experimental values obtained by simulation experiments. The approximation appears to be good. Finally, the huge differences in outcomes between Poisson and Erlang arrivals of the batches are highlighted.     

Controlling job arrivals with processing time windows into Batch Processor Buffer

We consider a two-stage manufacturing system composed of a batch processor and its upstream processor. Jobs exit the upstream processor and join a queue in front of the batch processor, where they wait to be processed. The batch processor has a finite capacity Q, and its processing time is independent of the number of jobs loaded into the batch processor. In certain manufacturing systems (including semiconductor wafer fabrication), a processing time window exists from the time instance the job exits the upstream processor till the time instance it enters the batch processor. If a job is not processed before reaching the end of its processing time window, job rework or validation is required. We model this drawback by assigning a reward R for each successfully processed job by the upstream processor, and a penalty C for each job that reaches the end of its processing time window without being processed by the batch processor. We initially assume an infinite job source in front of the serial processor and also assume that the batch processor is operated under a threshold policy. We provide a method for controlling the production of the serial processor, considering the processing time window between the upstream processor and the downstream batch processor. We then show how the serial processor control policy can be modified when the serial processor also experiences intermittent job arrival.     

考虑路径冲突的AGV配置与调度优化

合理调度有限的码头资源以满足船舶的装卸时间要求是自动化集装箱码头的重要目标之一。针对自动化集装箱码头自动导引车(automated guided vehicle,AGV)配置与调度问题,考虑船舶装卸时间要求和AGV运输过程中的路径冲突,提出分阶段调度策略。将船舶装卸作业分为卸船阶段、装卸同步阶段、装船阶段三个阶段,在每个阶段中,建立以最小化最大完工时间和最小化AGV空载和等待时间为双目标的调度优化模型,并设计基于NSGA-Ⅱ的启发式算法求解。根据本阶段的实际完工时间,从最优解集中选择下一阶段AGV的配置与调度方案。最后对比其他调度方案表明本文调度方案能够满足集装箱船的装卸时间要求,且提高了AGV的利用率,更符合码头实际作业要求。     

A spatiotemporal Data Envelopment Analysis (S-T DEA) approach: the need to assess evolving units

We consider a two-stage supply chain with one supplier and one manufacturer. The manufacturer faces a Poisson demand process where the arrival rate depends on the selling price, the announced delivery time, and the delivery reliability defined as the probability of satisfying the announced delivery time. Such a demand model generalizes the works in the literature by simultaneously considering the above three demand sensitivity factors. The main purpose of this paper is to study the equilibrium decisions in the supply chain with an all-unit quantity discount contract. We consider four scenarios regarding whether the leadtime standard, the delivery reliability standard, and the manufacturer’s capacity are endogenous, and whether the manufacturer’s production cost is its private information. We find that an all-unit quantity discount scheme can coordinate the supply chain for most cases. Managerial insights are observed regarding the impact of the three demand sensitivity factors. For example, the breakpoint in an optimal quantity discount contract always increases with the delivery reliability sensitivity under an exogenous delivery reliability, but may decrease under an endogenous delivery reliability; with asymmetric information, a higher variance of the manufacturer’s unit production costs leads to a lower unit wholesale price for the low-cost manufacturer.     

Single-machine scheduling of multi-operation jobs without missing operations to minimize the total completion time

We consider the problem of scheduling multi-operation jobs on a singe machine to minimize the total completion time. Each job consists of several operations that belong to different families. In a schedule each family of job operations may be processed as batches with each batch incurring a set-up time. A job is completed when all of its operations have been processed. We first show that the problem is strongly NP-hard even when the set-up times are common and each operation is not missing. When the operations have identical processing times and either the maximum set-up time is sufficiently small or the minimum set-up time is sufficiently large, the problem can be solved in polynomial time. We then consider the problem under the job-batch restriction in which the operations of each batch is partitioned into operation batches according to a partition of the jobs. We show that this case of the problem can be solved in polynomial time under a certain condition.     

可拆分平行机排序问题研究

平行机排序问题是把n个产品安排到m台机器上加工，使其总费用最小．通常的平行机排序问题都假设（C1）：任何产品不能在不同机器上同时加工．但是，如果把产品的加工时间看成一个产品量的需求，就可以假设（C2）：允许同一产品拆分在不同机器上同时加工．本文首先回顾了C1假设下平行机排序问题已有的结果，然后基于假设C2，讨论了各种费用目标下问题的算法及其复杂性．在没有生产准备时间的情况下，给出了一些问题的多项式算法和线性规划方法．在有独立生产准备时间的情况下，给出了P／split／Cmax问题的启发式算法及其算法分析．     

基于时间的供应链物流运作能力计划模型研究

本文考察了在基于时间的多阶供应链环境下的物流运作能力计划,主要的系统成本是物流要素能力固定成本和库存成本.为了决定最优的稳定周期性的物流运作能力计划,提出了一般性的整数非线性规划公式.强调了在这样的系统运作中的开始时间的关键作用,报告了在确保在任何阶段没有缺货条件下的物流能力决策的一些有价值的结论.