Erhan Kutanoglu

Integrating preventive maintenance planning and production scheduling for a single machine

Preventive maintenance planning, and production scheduling are two activities that are inter-dependent but most often performed independently. Considering that preventive maintenance, and repair affect both available production time, and the probability of machine failure, we are surprised that this inter-dependency seems to be overlooked in the literature. We propose an integrated model that coordinates preventive maintenance planning decisions with single-machine scheduling decisions so that the total expected weighted completion time of jobs is minimized. Note that the machine of interest is subject to minimal repair upon failure, and can be renewed by preventive maintenance. We investigate the value of integrating production scheduling with preventive maintenance planning by conducting an extensive experimental study using small scheduling problems. We compare the performance of the integrated solution with the solutions obtained from solving the preventive maintenance planning, and job scheduling problems independently. For the problems studied, integrating the two decision-making processes resulted in an average improvement of approximately 2% and occasional improvements of as much as 20%. Depending on the nature of the manufacturing system, an average savings of 2% may be significant. Certainly, savings in this range indicate that integrated preventive maintenance planning, and production scheduling should be focused on critical (bottleneck) machines. Because we use total enumeration to solve the integrated model for small problems, we propose a heuristic approach for solving larger problems. Our analysis is based on minimizing total weighted completion time; thus, both the scheduling, and maintenance problems favor processing shorter jobs in the beginning of the schedule. Given that due-date-based objectives, such as minimizing total weighted job tardiness, present more apparent trade-offs & conflicts between preventive maintenance planning, and job scheduling, we believe that integrated preventive maintenance planning & production scheduling is a worthwhile area of study.

Empty container management for intermodal transportation networks

Abstract We present a computational analysis of the effect of planning horizon length on empty container management for intermodal transportation networks. The analysis is based on an integer program that seeks to minimize total costs related to moving empty containers, subject to meeting requirements for moving loaded containers. A case study of potential container-on-barge operations within the Mississippi River basin illustrates the effects of planning horizon length on mode selection. A longer planning horizon can encourage the use of inexpensive, slow transportation modes, such as barge. The impact depends on the number and location of container storage pools.

On combinatorial auction and Lagrangean relaxation for distributed resource scheduling

Most existing methods for scheduling are based on centralized or hierarchical decision making using monolithic models. In this study, we investigate a new method based on a distributed and locally autonomous decision structure using the notion of combinatorial auction. In combinatorial auction the bidders demand a combination of dependent objects with a single bid. We show that not only can we use this auction mechanism to handle complex resource scheduling problems, but there exist strong links between combinatorial auction and Lagrangean-based decomposition. Exploring some of these properties, we characterize combinatorial auction using auction protocols and payment functions. This study is a first step toward developing a distributed scheduling framework that maintains system-wide performance while accommodating local preferences and objectives. We provide some insights to this framework by demonstrating four different versions of the auction mechanism using job shop scheduling problems.

Energy-Aware Cooperative Content Distribution Over Wireless Networks: Optimized and Distributed Approaches

In this paper, we address the problem of optimal energy-aware content distribution over wireless networks with mobile-to-mobile cooperation. Given a number of mobile terminals (MTs) interested in downloading a common content via a base station (BS), the MTs are grouped into cooperative groups or coalitions. Within each coalition, an optimally chosen coalition head downloads the content from the BS and either unicasts or multicasts it to the other MTs. The centralized optimization formulations are derived for both unicasting and multicasting among the MTs, along with the suitable simplifications to reduce the complexity of the optimization formulations. Then, a polynomial time heuristic algorithm is proposed to solve the optimization problems for relatively large networks where the optimal solution becomes computationally complex. Furthermore, a distributed algorithm, which is based on coalitional game theory, is developed to allow the MTs to choose, independently, which coalitions to join. Performance results for various scenarios demonstrate that the proposed algorithms lead to significant reduction in the total energy consumed by the MTs. In addition, the proposed centralized and distributed algorithms are shown to have relatively low complexity while achieving a near-optimal performance.

Heuristic based scheduling system for diffusion in semiconductor manufacturing

This paper addresses the scheduling of lots in a specific wafer fabrication area, called diffusion, where scheduling of lots interact with batching, equipment dedication and queue time constraints. Realizing the difficulty of solving the underlying mathematical program optimally, we develop a heuristic to regularly schedule the lots available in the area in real time. The paper explains the user interface and implementation issues as well as the details of the heuristic logic. The results obtained from production in a wafer fabrication facility to date show high user compliance, improved predictability and visibility of the overall schedule, and improved operational performance including reduced cycle times and queue time violations.

Improving scheduling robustness via preprocessing and dynamic adaptation

We study methods to improve scheduling robustness under processing time variation. We propose a two-stage scheme that preprocesses the scheduling data to create a skeleton of a schedule and then completes it over time through dynamic adaptation. We focus our attention on the classical job shop scheduling problem. Preprocessing starts at the beginning of the planning period (at the time of scheduling) when a priori information becomes available on processing time uncertainty. We first decompose the job shop scheduling problem into network-structured subproblems using Lagrangian relaxation. For each subproblem, we introduce stochastic constraints that capture the processing time uncertainty. We incorporate the stochastic information in such a way that the subproblems retain their efficient network structure. Using a subgradient search algorithm, we iteratively improve the lower and upper bounds obtained from the Lagrangian relaxed problem, which produce a partial sequence of critical operations. This so-called Lagrangian ranking defines a preprocessed schedule where the complete scheduling is determined dynamically over time, adapting to changing shop conditions. We test the robustness of the two-stage scheme through extensive computational experiments. We show that the scheme significantly outperforms deterministic scheduling methods and dynamic dispatching rules with minimal computational burden.

Production Planning and Scheduling: Interaction and Coordination

In many organizations, production planning is part of a hierarchical planning, capacity/resource allocation, scheduling and control framework. The production plan considers resource capacities, time periods, supply and demand over a reasonably long planning horizon at a high level. Its decision then forms the input to the more detailed, shorter-term functions such as scheduling and control at the lower level, which usually have more accurate estimates of supply, demand, and capacity levels. Hence, interaction between production planning and production scheduling/control is inevitable, not only because the scheduling/control decisions are constrained by the planning decisions, but also because disruptions occurring in the execution/control stage (usually after schedule generation) may affect the optimality and/or feasibility of both the plan and the schedule. If the overall performance of the production system is to be improved, disruptions must be managed effectively, with careful consideration of both planning and scheduling decisions. This chapter focuses on the interaction between production planning and scheduling, emphasizing the coordination of decisions, with special emphasis on making robust decisions at both levels in the face of unexpected disruptions. We provide examples and realistic scenarios from semiconductor manufacturing.

ELECTRICAL PARAMETER CONTROL FOR SEMICONDUCTOR DEVICE MANUFACTURING: A FABWIDE APPROACH

Abstract Wafers that fail to meet their electrical specifications lead to scrap which negatively impacts yield and manufacturing costs. Most existing research has focused on controlling individual steps during the manufacturing process via run-to-run control, but almost no work has looked at directly controlling the electrical characteristics. A control scheme is proposed to directly control electrical parameter values. The control algorithm uses a model to predict electrical parameter values after each processing step and determines optimal adjustments for the future processing steps. Simulation results show significant reduction in electrical parameter variations for both constrained and unconstrained control.

Reactive Tabu Search for Large Scale Service Parts Logistics Network Design and Inventory Problems

This chapter documents a study of a reactive tabu search (RTS) approach to the integrated service part logistics (SPL) network design and inventory stocking problem. The integrated problem of designing and stocking an SPL network has attracted more attention recently. The two sets of decisions (network design and inventory stocking) usually have been considered separately and sequentially in practice as well as in the research literature, although interdependency between them exists and integration is necessary for overall system performance optimization. However, the integrated mathematical programming model and solution development for this problem are often intractable due to the time-based service constraints which confine the lower bound of the demand percentage satisfied within the specified time windows. We use a RTS method to efficiently find very good solutions to this problem. Tabu search combines a hill climbing strategy with a memory structure which guides the search. The reactive mechanism dynamically adjusts the tabu tenure during the search. An escape mechanism is activated when the search is trapped in a local attractor basin. We also apply heuristic techniques to construct the initial solution and use rule based comparisons to determine the best non-tabu solution in a neighborhood about the current incumbent solution. By applying this metaheuristic method to problem sets of different sizes, we obtain high-quality solutions with remarkably small amounts of computational effort. For the smaller problems, the tabu search solution is identical or very close to the optimal solution provided by classical optimization-based methods. For the larger problems, RTS obtains solutions superior to those obtained by classical approaches.