Research in the laboratory focuses on a wide range of logistics problems in manufacturing, supply chains, transportation and service systems. A common thread to research in these areas is the use of operations research methods including stochastic modeling and optimization. The following is a sample of current research projects.

Contract Manufacturing in the Electronic Industry

Principal Investigators:

  • Saif Benjaafar, University of Minnesota
  • Karen Donohue, University of Minnesota
  • David Wu, Lehigh University
  • N. Vishwanadham, National University of Singapore
  • Fikri Karaesmen, Ecole Centrale de Paris

Sponsors: The Logistics Institute-Asia Pacific, Polarfab Corporation, Lucent Technology

Contract Manufacturing has emerged in recent years as the preferred mode of production in the high-tech industry. Contract manufacturing has grown from a few billion dollars in the early 1990s to over a $140 billion industry in 2000. By 2005, over 50% of all electronics manufacturing is expected to be carried out on a contract-basis. Contract manufacturing is different from traditional supplier-manufacturer settings in that few large supercontractors dominate the marketplace. These supercontractors are able to support a large number of original equipment manufacturers (OEMs) with products ranging from cellular phones to laser printers. In this sense, supercontractors are engaged less in the selling of specific products and more in the selling of manufacturing capacity. Contract manufacturing raises several important research questions. How should contractual agreements between contractors and OEMs be structured so that they are mutually beneficial? How should contractors ration their available capacity and inventory among the competing needs of different customers? Should contract manufacturers pursue a few large accounts or several smaller ones? What are the true costs and benefits of customer diversification? When there is excess capacity, how should it be channeled? How should contractors take advantage of the emerging e-markets for capacity? What are the costs and benefits of the dynamic pricing of capacity that results from participating in these marketplaces?

In this research, we propose to address these important questions. More specifically, we will develop analytical models, computational algorithms, and decision support tools that push the current frontier of supply chain modeling and analysis in this area. These tools will assist contract manufacturers in managing their capacity, customer and product portfolios, and inventories and provide them with vital insights as they face the new challenges of supercontracting. Our research will focus on three main areas:

  1. Contract Design, where we will develop a game-theoretic framework for the modeling and analysis of option-based contracts in multi-buyer settings. Various forms of contracts will be examined and their implications to supercontracting will be analyzed.
  2. Variety Management, where we will develop analytical models that draw from both inventory and queueing theory to assist contractors in making decisions regarding product and customer portfolio selection, dynamic capacity allocation, and inventory rationing.
  3. Capacity Trading, where we will examine game-theoretic models for the emerging electronic marketplaces for capacity. In particular, we will examine how participation in online markets such as auctions affects the way contract manufacturers price and manage their capacity.

Our overarching objective is to develop a science base for the emerging field of contract manufacturing. We will introduce new research issues brought forward by this fast growing sector of the economy, and the role various analytical models play in providing vital insights. Our research will benefit industry by providing decision-makers with analytical tools that can guide both their contractual and operational decisions. Our research will also provide a roadmap to contract manufacturers regarding when and how to participate in the new e-markets for manufacturing capacity. In all phases of this research, we will be guided by continuous interaction with our industry partners.

Product Design and Inventory Deployment for Improving Delivery Time Performance in the Steel Industry

Principal investigators:

  • Diwakar Gupta and Saifallah Benjaafar, University of Minnesota

Sponsors: NSF

This grant provides funding for the development of mathematical models and numerical tools which can be used to determine the optimal design (dimensions, grade, and weight) of steel slabs, and optimal replenishment policies for slab inventories. A two-step approach will be used. In the first step, optimal design configurations will be determined using a combination of heuristics to generate both a good initial solution and refinements to the initial solution. A stochastic integer programming formulation involving binary variables will be used to provide bounds on the performance of these heuristic solutions. For a given number of slab designs, these methods will determine those configurations that maximize coverage, measured in terms of total finished tons that can be manufactured from the chosen configurations. In the second step, a multi period stochastic linear program will be developed to determine replenishment batch sizes for each slab design configuration in order to minimize the sum of inventory holding and production inefficiency costs.

If successful, the results of this research will provide a science-based solution to a chronic problem faced by integrated steel mills (ISMs). By focusing on reducing the number of slab designs, ISMs can pursue specialty steel markets while simultaneously reducing their production process complexity and inventory costs. Solution techniques developed to solve the underlying stochastic optimization problems, will have wider applications to industries with similar process architecture, e.g., paper and pulp, and to other stochastic optimization problems involving a large number of scenarios. The latter occur in a host of applications ranging from energy models, capacity planning to financial asset management. Models proposed for determining optimal replenishment policies, especially when coupled with the possibility of delaying product differentiation, will be useful for manufacturing firms trying to cope with increased product variety.

Design of Production-Inventory Systems

Principal investigators:

  • Saif Benjaafar and Bill Cooper, University of Minnesota
  • Mohsen Elhafsi, University of California, Riverside

Sponsor: NSF, St Jude Medical

The focus of this research is on modeling, design and analysis of integrated production-inventory systems. Several projects are underway, including:

  1. Analysis of inventory pooling in production-inventory systems
  2. Demand allocation in multi-product/multi-facility make-to-stock production systems
  3. Analysis of demand variability in production-inventory systems
  4. Inventory rationing in make-to-stock systems
  5. Advanced order information in production-inventory systems.

Flexible Queueing Systems

Principal investigators:

  • Saif Benjaafar, University of Minnesota
  • Gurumurthi Suryanarayan, Aloca, Corporations

This research deals with the modeling and analysis of flexible queueing systems. Flexible queueing systems to refer to systems with multiple classes of arrivals and multiple servers, where customers have the flexibility of being routed to more than one server and servers possess the capability of processing more than one customer class. Flexible queueing systems arise in a variety of contexts, including manufacturing, telecommunication networks, computer systems, vehicle and crew dispatching, and service systems, among others. We provide a modeling framework for the analysis of general system with an arbitrary number of heterogeneous resources and customer types and arbitrary routing and resource flexibility. We consider a rich set of control policies that include strict priority schemes for customer routing and queue selection. We use our models to generate several insights into the effect of system parameters. In particular, we examine the relationship between flexibility, control policies, and system throughput.