Stochastic multi-site capacity planning of TFT-LCD manufacturing using expected shadow-price based decomposition

This paper presents a stochastic optimization model and efficient decomposition algorithm for multi-site capacity planning under the uncertainty of the TFT-LCD industry. The objective of the stochastic capacity planning is to determine a robust capacity allocation and expansion policy hedged against demand uncertainties because the demand forecasts faced by TFT-LCD manufacturers are usually inaccurate and vary rapidly over time. A two-stage scenario-based stochastic mixed integer programming model that extends the deterministic multi-site capacity planning model proposed by Chen et al. (2010) [1] is developed to discuss the multi-site capacity planning problem in the face of uncertain demands. In addition a three-step methodology is proposed to generate discrete demand scenarios within the stochastic optimization model by approximating the stochastic continuous demand process fitted from the historical data. An expected shadow-price based decomposition, a novel algorithm for the stage decomposition approach, is developed to obtain a near-optimal solution efficiently through iterative procedures and parallel computing. Preliminary computational study shows that the proposed decomposition algorithm successfully addresses the large-scale stochastic capacity planning model in terms of solution quality and computation time. The proposed algorithm also outperforms the plain use of the CPLEX MIP solver as the problem size becomes larger and the number of demand scenarios increases.     

Stochastic binary problems with simple penalties for capacity constraints violations

This paper studies stochastic programs with first-stage binary variables and capacity constraints, using simple penalties for capacities violations. In particular, we take a closer look at the knapsack problem with weights and capacity following independent random variables and prove that the problem is weakly ${\mathcal{N}\mathcal{P}}$ -hard in general. We provide pseudo-polynomial algorithms for three special cases of the problem: constant weights and capacity uniformly distributed, subset sum with Gaussian weights and strictly positively distributed random capacity, and subset sum with constant weights and arbitrary random capacity. We then turn to a branch-and-cut algorithm based on the outer approximation of the objective function. We provide computational results for the stochastic knapsack problem (i) with Gaussian weights and constant capacity and (ii) with constant weights and capacity uniformly distributed, on randomly generated instances inspired by computational results for the knapsack problem.     

A stochastic programming approach for planning horizons of infinite horizon capacity planning problems

Planning horizon is a key issue in production planning. Different from previous approaches based on Markov Decision Processes, we study the planning horizon of capacity planning problems within the framework of stochastic programming. We first consider an infinite horizon stochastic capacity planning model involving a single resource, linear cost structure, and discrete distributions for general stochastic cost and demand data (non-Markovian and non-stationary). We give sufficient conditions for the existence of an optimal solution. Furthermore, we study the monotonicity property of the finite horizon approximation of the original problem. We show that, the optimal objective value and solution of the finite horizon approximation problem will converge to the optimal objective value and solution of the infinite horizon problem, when the time horizon goes to infinity. These convergence results, together with the integrality of decision variables, imply the existence of a planning horizon. We also develop a useful formula to calculate an upper bound on the planning horizon. Then by decomposition, we show the existence of a planning horizon for a class of very general stochastic capacity planning problems, which have complicated decision structure.     

Stochastic programming based capacity planning for semiconductor wafer fab with uncertain demand and capacity

Capacity planning is a challenging problem in semiconductor manufacturing industry due to high uncertainties both in market and manufacturing systems, short product life cycle, and expensive capital invest. To tackle this problem, this paper proposes a scenario-based stochastic programming model which considers demand and capacity uncertainties via scenarios, where the overall equipment efficiency is employed to describe the uncertain capacity for the first time. Based on the decentralized structure of tool procurement, production, stockout, and inventory decision-making processes, recourse approximation strategies are presented with varying degree of information share. The computational experiments show that the resulting tool set is robust enough to cope with the changes in capacity with the expected profits being maximized for different scenarios, and the scheme can generate pretty good solutions in reasonable computational time.     

Supply capacity acquisition and allocation with uncertain customer demands

We study a class of capacity acquisition and assignment problems with stochastic customer demands often found in operations planning contexts. In this setting, a supplier utilizes a set of distinct facilities to satisfy the demands of different customers or markets. Our model simultaneously assigns customers to each facility and determines the best capacity level to operate or install at each facility. We propose a branch-and-price solution approach for this new class of stochastic assignment and capacity planning problems. For problem instances in which capacity levels must fall between some pre-specified limits, we offer a tailored solution approach that reduces solution time by nearly 80% over an alternative approach using a combination of commercial nonlinear optimization solvers. We have also developed a heuristic solution approach that consistently provides optimal or near-optimal solutions, where solutions within 0.01% of optimality are found on average without requiring a nonlinear optimization solver.     

Models for single-sector stochastic air traffic flow management under reduced airspace capacity

Air traffic efficiency is heavily influenced by unanticipated factors that result in capacity reduction. Of these factors, weather is the most significant cause of delays in airport and airspace operations. Considering weather-related uncertainty, air traffic flow management involves controlling air traffic through allocation of available airspace capacity to flights. The corresponding decision process results in a stochastic dynamic problem where aircraft on the ground and in the air are controlled based on the evolution of weather uncertainty. We focus on the single-sector version of the problem that is applicable to a majority of cases where a volume of airspace has reduced capacity due to convective weather. We model the decision process through stochastic integer programming formulations and computationally analyse it for tractability. We then demonstrate through actual flight schedule data that a simplistic but practically implementable approximation procedure is a generally effective solution approach for these models.     

Integrated capacity and inventory management with capacity acquisition lead times

We model a make-to-stock production system that utilizes permanent and contingent capacity to meet non-stationary stochastic demand, where a constant lead time is associated with the acquisition of contingent capacity. We determine the structure of the optimal solution concerning both the operational decisions of integrated inventory and flexible capacity management, and the tactical decision of determining the optimal permanent capacity level. Furthermore, we show that the inventory (either before or after production), the pipeline contingent capacity, the contingent capacity to be ordered, and the permanent capacity are economic substitutes. We also show that the stochastic demand variable and the optimal contingent capacity acquisition decisions are economic complements. Finally, we perform numerical experiments to evaluate the value of utilizing contingent capacity and to study the effects of capacity acquisition lead time, providing useful managerial insights.     

Base stock policies for production/inventory problems with uncertain capacity levels

In this paper we consider a single item, stochastic demand production/inventory problem where the maximum amount that can be produced (or ordered) in any given period is assumed to be uncertain. Inventory levels are reviewed periodically. The system operates under a stationary modified base stock policy. The intent of our paper is to present a procedure for computing the optimal base stocl level of this policy under expected average cost per period criterion. This procedure would provide guidance as to the appropriate amount of capacity to store in the form of inventory in the face of stochastic demand and uncertain capacity. In achieving this goal, our main contribution is to establish the analogy between the class of base stock production/inventory policies that operate under demand/capacity uncertainty, and the G/G/1 queues and their associated random walks. We also present example derivations for some important capacity distributions.     

Inventory control with Markovian capacity and the option of order rejection

We consider a stochastic inventory control problem with Markovian capacity and the option of order rejection. We show its optimal policy to be the combination of a capacity-dependent modified base-stock production policy and a capacity-dependent critical-point order acceptance policy. When capacity is stochastically monotone, we show that the policy parameters change over the current capacity level in an intuitive way. All these results can be extended to the infinite-horizon case, with or without cost discounting. Our computational study substantiates the benefits from both exploiting the Markovian behavior of the capacity and rejecting orders when necessary.     

Managing capacity flexibility in make-to-order production environments

This paper addresses the problem of managing flexible production capacity in a make-to-order (MTO) manufacturing environment. We present a multi-period capacity management model where we distinguish between process flexibility (the ability to produce multiple products on multiple production lines) and operational flexibility (the ability to dynamically change capacity allocations among different product families over time). For operational flexibility, we consider two polices: a fixed allocation policy where the capacity allocations are fixed throughout the planning horizon and a dynamic allocation policy where the capacity allocations change from period to period. The former approach is modeled as a single-stage stochastic program and solved using a cutting-plane method. The latter approach is modeled as a multi-stage stochastic program and a sampling-based decomposition method is presented to identify a feasible policy and assess the quality of that policy. A computational experiment quantifies the benefits of operational flexibility and demonstrates that it is most beneficial when the demand and capacity are well-balanced and the demand variability is high. Additionally, our results reveal that myopic operating policies may lead a firm to adopt more process flexibility and form denser flexibility configuration chains. That is, process flexibility may be over-valued in the literature since it is assumed that a firm will operate optimally after the process flexibility decision. We also show that the value of process flexibility increases with the number of periods in the planning horizon if an optimal operating policy is employed. This result is reversed if a myopic allocation policy is adopted instead.     

Joint replenishment and pricing decisions in inventory systems with stochastically dependent supply capacity

In this paper, we study a joint optimization problem of replenishment and pricing for a periodic-review inventory system with random supply capacity. When making replenishment and pricing decisions at the beginning of each period, the firm only knows the supplier’s available capacity in the current period, but does not know what will be the available capacity in future periods. The salient feature of our model is that the random supply capacities for different periods are dependent. Several stochastic dependency structures are considered for the supply capacity sequence, including the one-lag and the multi-lag dependency.     

Wireless network capacity management: A real options approach

This paper applies financial option valuation methods to new wireless network capacity investment decision timing. In particular, we consider the case of network capacity for cellular telephone service. Given a cluster of base stations (with a certain traffic capacity per base station), we determine when it is optimal to increase capacity for each of the base stations contained in the cluster. We express this in terms of the fraction of total cluster capacity in use, i.e. we calculate the optimal time to upgrade in terms of the ratio of observed usage to existing capacity. We study the optimal decision problem of adding new capacity in the presence of stochastic wireless demand for services. A four factor algorithm is developed, based on a real options formulation. Numerical examples are provided to illustrate various aspects of the model.     

Designing multi-echelon service parts networks with finite repair capacity

We propose an approach to model and solve the joint problem of facility location, inventory allocation and capacity investment in a two echelon, single-item, service parts supply chain with stochastic demand. The objective of the decision problem is to minimize the total expected costs associated with (1) opening repair facilities, (2) assigning each field service location to an opened facility, (3) determining capacity levels of the opened repair facilities, and (4) optimizing inventory allocation among the locations. Due to the size of the problem, computational efficiency is essential. The accuracy of the approximations and effectiveness of the approach are analyzed with two numerical studies. The approach provides optimal results in 90% of scenarios tested and was within 2% of optimal when it did not.We explore the impact of capacity utilization, inventory availability, and lead times on the performance of the approach. We show that including tactical considerations jointly with strategic network design resulted in additional cost savings from 3% to 12%. Our contribution is the development of a practical model and approach to support the decision making process of joint facility location and multi-echelon inventory optimization.     

Managing manufacturing risks by using capacity options

In this study, we investigate the strategy of increasing production capacity temporarily through contingent contractual agreements with short-cycle manufacturers to manage the risks associated with demand volatility. We view all these agreements as capacity options. More specifically, we consider a manufacturing company that produces a replenishment product that is sold at a retailer. The demand for the product switches randomly between a high level and a low level. The production system has enough capacity to meet the demand in the long run. However, when the demand is high, it does not have enough capacity to meet the instantaneous demand and thus has to produce to stock in advance. Alternatively, a contractual agreement with a short-cycle manufacturer can be made. This option gives the right to receive additional production capacity when needed. There is a fixed cost to purchase this option for a period of time and, if the option is exercised, there is an additional per unit exercise price which corresponds to the cost of the goods produced at the short-cycle manufacturer. We formulate the problem as a stochastic optimal control problem and analyse it analytically. By comparing the costs between two cases where the contract with the short-cycle manufacturer is used or not, the value of this option is evaluated. Furthermore, the effect of demand variability on this contract is investigated.     

Approximate dynamic programming for capacity allocation in the service industry

We consider a problem where different classes of customers can book different types of service in advance and the service company has to respond immediately to the booking request confirming or rejecting it. The objective of the service company is to maximize profit made of class-type specific revenues, refunds for cancellations or no-shows as well as cost of overtime. For the calculation of the latter, information on the underlying appointment schedule is required. In contrast to most models in the literature we assume that the service time of clients is stochastic and that clients might be unpunctual. Throughout the paper we will relate the problem to capacity allocation in radiology services. The problem is modeled as a continuous-time Markov decision process and solved using simulation-based approximate dynamic programming (ADP) combined with a discrete event simulation of the service period. We employ an adapted heuristic ADP algorithm from the literature and investigate on the benefits of applying ADP to this type of problem. First, we study a simplified problem with deterministic service times and punctual arrival of clients and compare the solution from the ADP algorithm to the optimal solution. We find that the heuristic ADP algorithm performs very well in terms of objective function value, solution time, and memory requirements. Second, we study the problem with stochastic service times and unpunctuality. It is then shown that the resulting policy constitutes a large improvement over an “optimal” policy that is deduced using restrictive, simplifying assumptions.     

Resource capacity allocation to stochastic dynamic competitors: knapsack problem for perishable items and index-knapsack heuristic

In this paper we propose an approach for solving problems of optimal resource capacity allocation to a collection of stochastic dynamic competitors. In particular, we introduce the knapsack problem for perishable items, which concerns the optimal dynamic allocation of a limited knapsack to a collection of perishable or non-perishable items. We formulate the problem in the framework of Markov decision processes, we relax and decompose it, and we design a novel index-knapsack heuristic which generalizes the index rule and it is optimal in some specific instances. Such a heuristic bridges the gap between static/deterministic optimization and dynamic/stochastic optimization by stressing the connection between the classic knapsack problem and dynamic resource allocation. The performance of the proposed heuristic is evaluated in a systematic computational study, showing an exceptional near-optimality and a significant superiority over the index rule and over the benchmark earlier-deadline-first policy. Finally we extend our results to several related revenue management problems.     

Single-pass and approximate dynamic-programming algorithms for order acceptance and capacity planning

This paper investigates dynamic order acceptance and capacity planning under limited regular and non-regular resources. Our goal is to maximize the profits of the accepted projects within a finite planning horizon. The way in which the projects are planned affects their payout time and, as a consequence, the reinvestment revenues as well as the available capacity for future arriving projects. In general, project proposals arise dynamically to the organization, and their actual characteristics are only revealed upon arrival. Dynamic solution approaches are therefore most likely to obtain good results. Although the problem can theoretically be solved to optimality as a stochastic dynamic program, real-life problem instances are too difficult to be solved exactly within a reasonable amount of time. Efficient and effective heuristics are thus required that supply a response without delay. For this reason, this paper considers both ‘single-pass’ algorithms as well as approximate dynamic-programming algorithms and investigates their suitability to solve the problem. Simulation experiments compare the performance of our procedures to a first-come, first-served policy that is commonly used in practice.     

An application of Lagrangian relaxation to a capacity planning problem under uncertainty

A supply chain network-planning problem is presented as a two-stage resource allocation model with 0-1 discrete variables. In contrast to the deterministic mathematical programming approach, we use scenarios, to represent the uncertainties in demand. This formulation leads to a very large scale mixed integer-programming problem which is intractable. We apply Lagrangian relaxation and its corresponding decomposition of the initial problem in a novel way, whereby the Lagrangian relaxation is reinterpreted as a column generator and the integer feasible solutions are used to approximate the given problem. This approach addresses two closely related problems of scenario analysis and two-stage stochastic programs. Computational solutions for large data instances of these problems are carried out successfully and their solutions analysed and reported. The model and the solution system have been applied to study supply chain capacity investment and planning.     

Two-resource stochastic capacity planning employing a Bayesian methodology

We examine a stochastic capacity-planning problem with two resources that can satisfy demand for two services. One of the resources can only satisfy demand for a specific service, whereas the other resource can provide both services. We formulate the problem of choosing the capacity levels of each resource to maximize expected profits. In addition, we provide analytic, easy-to-interpret optimal solutions, as well as perform a comparative statics analysis. As applying the optimal solutions effectively requires good estimates of the unknown demand parameters, we also examine Bayesian estimates of the demand parameters derived via a class of conjugate priors. We compare the optimal expected profits when demands for the two services follow independent distributions with informative and non-informative priors, and demonstrate that using good informative priors on demand can significantly improve performance.