Research.

These challenging economic times call for innovative thinking that could generate big productivity gains. This month in the Design and Manufacturing segment of IIE Transactions (Volume 41, No. 3), we highlight two articles that document benefits from research that has questioned longstanding assumptions regarding warehouse layout design and assembly line balancing that produce many product variants.

[ILLUSTRATION OMITTED]

New warehouse aisle designs

There are more than 500,000 ware houses in the United States and many more throughout the world. The simplest type of warehouse is called a "unit load warehouse" because it receives and ships goods typically in pallet quantities without intermediate operations that alter the unit load. The basic operations of a unit load ware house are to receive goods, store them until needed and then retrieve them from storage and load them for shipment.

Until recently, all known warehouses adhered to two implicit design rules: all rows of storage, and therefore also the aisles (called "picking aisles") between them, must be parallel to one another, and any cross aisles must be perpendicular to the picking aisles. Associate professor Kevin Gue of Auburn University and professor Russell Meller of the University of Arkansas asked, "Must warehouses follow these rules?" If not, what can be gained by new aisle layouts? The answer turns out to be quite a lot.

In "Aisle Configurations for Unit Load Warehouses," Gue and Meller present an optimization-based model for two new aisle layouts for unit load warehouses. The first is designed for situations in which all picking aisles are parallel to one another. Rather than inserting a straight cross aisle that is perpendicular to the picking aisles, the authors solve an optimization problem to determine the best cross aisle profile--a slightly curved "V" with the vertex of the V at the pickup and deposit point. Numerical results indicate that this "Flying-V" cross aisle reduces the average travel distance in a unit load warehouse by 8 percent to 10 percent.

The second aisle design relaxes the "rule" that all picking aisles must be parallel to one another. In this second design, a V-shaped cross aisle is again inserted into the picking space, but unlike the picking aisles above the V, which are vertical with respect to the upright V, the picking aisles on the lower left and lower right of the V are rotated 90 degrees so they are horizontal with respect to the V. This forms what the authors call the "fishbone aisle design," which offers the advantage that every trip in the warehouse uses a shortcut on the cross aisle and therefore never traverses the full distance on an inefficient "Manhattan grid." As a result, the average travel distance declines by up to 20 percent.

Using the results of this research, two warehouses in the U.S. have implemented designs that do not conform to the traditional standards. At both facilities, management and workers are happy with the designs and the performance improvements attributed to them. Perhaps the design rules are starting to change.

CONTACT: Kevin Gue; krgue@eng.auburn.edu; (334) 844-1425; Department of Industrial and Systems Engineering, Auburn University, Auburn, AL 36849; and Russell Meller; rmeller@uark.edu; (479) 575-6196; Department of Industrial .Engineering, University of Arkansas, Fayetteville, AR 72701

Balancing high-variety assembly lines

Even for the Model T Ford, balancing the assembly line--assigning tasks to assembly stations to equalize the workloads--was a challenging problem. Today, when a single vehicle model might be produced in millions of variations, and a single assembly line may produce a few different vehicle models, the task is even more daunting. If assembly lines are not well-balanced, capital equipment and labor are not utilized as efficiently as they could be. Among the constraints that must be satisfied when solving an assembly line balancing (ALB) problem are precedence constraints that specify which tasks must be performed before each of the others. Simply capturing this information is a formidable task.

[ILLUSTRATION OMITTED]

Professors Nils Boysen and Armin Scholl as well as research fellow Malte Fliedner of Friedrich-Schiller University of Jena, Germany, were motivated to study assembly line balancing for high-variety production lines after observing how it was performed at a German automobile manufacturer. They noticed that the approaches used at this company focused on so-called product options, such as choices among engine sizes or the presence or absence of certain features such as a sunroof. This general approach, however, ignores the possibility that a task may be associated with a combination of options, not just one. The simplified methods in use dramatically reduce the effort required for data collection and forecasting, but they lead to an overestimate of labor requirements, especially for tasks involving electronic components, where a single electronic component or module may support several different options.

Although managers and engineers at the automobile company were aware of this phenomenon, they did not regard it as a significant factor in line balancing. To understand the impact of this phenomenon, the authors developed a methodology that identifies the required information base and facilitates a more accurate representation of precedence constraints when tasks may be associated with combinations of options. With this, the authors are able to obtain improved assembly line balancing solutions. Computational experiments show that even a moderate number of multiple-option tasks may have a substantial effect on the quality of line balances. This finding raised the awareness of the planning team at the automobile company regarding the importance of properly accounting for such multi-option tasks, which would enable a reduction in costs that is so important in an industry where profit margins are already thin and under increasing pressure.

CONTACT: Armin Scholl; a.scholl@wiwi.uni-jena.de; + 049 3641/9 43170; Faculty of Economics and Business Administration, Friedrich-Schiller University of Jena, Carl-Zeiss-Strasse 3, 07743 Jena, Germany

The most recent issue of The Engineering Economist (Volume 53, No. 4) is focused on energy economics with Thomas Boucher, professor of industrial engineering at Rutgers University, serving as guest editor. One article features strategies for trading electricity and another examines complex decisions in a network, such as capacity replacement decisions in an electric power network.

Improved electricity trading strategies

As the regulation of electric utilities continues to lessen, trading electricity has become as commonplace as trading stocks on the New York Stock Exchange. In order to take advantage of price swings throughout the country, a power producer may sell its excess capacity (or even everyday capacity) in order to generate additional profits. This has occurred through the formation of day-ahead and real-time markets operated by independent system operators.

The risk in selling electricity a day ahead exposes a utility to forced outage risks because a utility is committed to providing the electricity. If a utility cannot deliver the electricity in real-time due to forced outages, it must purchase electricity in a possibly highly volatile real-time market in order to make up for the shortfall.

While electricity trading has been examined in the literature for years, it has generally looked at longer-term trading strategies, especially when considering capacity expansion decisions (such as when to build a new power generation plant). In "Short-Term Electric Power Trading Strategies for Portfolio Optimization," the authors examine day-ahead and real-time electricity trading strategies in order to maximize profits while limiting risk exposure.

As the prices are stochastic, a simulation approach is used to solve the model in the given framework. Numerical examples show that the proposed strategy improves the performance of a power portfolio over those strategies commonly used by traders.

CONTACT: K. Jo Min, Industrial and Manufacturing Systems Engineering, Iowa State University, Ames, IA, 50011

interconnected decisions

As engineered systems become more complex, decisions regarding their construction, maintenance, upgrading and replacement have become more complex, with many stakeholders who often have different objectives. In "Electric Power Network Decision Effects," an analysis is presented where local plant managers must decide when to replace existing power generation equipment (coal, oil, natural gas, nuclear and hydropower) with an equivalent number of 12-megawatt natural gas-fueled micro-turbine generators. Competing objectives include cost, reliability and environmental impact.

The paper actually takes a much wider scope than electricity networks, looking at interconnected decisions in general. The paper first reviews relevant factors in decision making, then proposes that in a network, some decisions can be considered individually while others should be considered plurally.

As it may not be feasible, or reasonable, to consider the impact of one's decision on all other entities in the network, a systems approach is needed to determine when decisions should be made jointly versus alone. The authors define this movement from a small decision, which can be made alone because the impact on others is minimal, to a big decision, which should be made as a group to minimize impact, according to the "size" of the decision. The electricity network example is used to illustrate this point.

CONTACT: Deborah L. Thurston, Department of Industrial and Enterprise Systems Engineering, University of Illinois at Urbana-Champaign, Urbana, IL 61801

Print Get the Mag Weekly Updates