Home | Business News | Browse by Publication | I | Interfaces

Dell uses a new production-scheduling algorithm to accommodate increased product variety.

Publication: Interfaces
Publication Date: 01-MAY-07
Format: Online
Delivery: Immediate Online Access

Article Excerpt
Early in 2003, with continually increasing product variety and production volumes, Dell, Inc. had reached the designed capacity limits of one of its main production facilities, the Morton L. Topfer Manufacturing Center (TMC). In 2004, TMC was facing a doubling of the number of product families it produced, with an anticipated degradation In production rates of nearly 20 percent. To help assuage this problem, we developed a new production-scheduling algorithm, which contains both optimization and heuristic components. The algorithm schedules product families on parallel, identical kitting lines to minimize the number of setups required and to reduce downtime and slow time during setups. Because of our work, Dell was able to accommodate the twofold increase in product variety, as well as an effective production-volume increase of over 35 percent. Furthermore, Dell realized a conservative cost avoidance of more than $1 million annually, primarily because it saved overtime costs that it would have required, in the absence of our solution, to handle the increases in production volume and product variety. This solution has been in operation at TMC since June 2004.

Key words: production scheduling; heuristic; optimization; product families; parallel lines; assemble-to-order.

History: This paper was refereed.

**********

This paper describes a production-scheduling problem in the manufacturing process of a high-volume, assemble-to-order, electronics manufacturer, and presents a scheduling algorithm to solve this problem. The problem arose in Dell's Morton L. Topfer Manufacturing Center (TMC) in Austin, Texas. Dell's direct, assemble-to-order business model allows it to do business directly with the customer and to assemble products only after a customer places an order (Dell and Magretta 1998); Dell designed TMC to support this model. Opened in 2001, TMC is both flexible and capable of producing a large variety of products. Given the typical trade-off between production volume and product variety (Stevenson 2001), this is an atypical situation. This combination of flexibility and high-volume production gives rise to the interesting nature of the scheduling problem we present.

This paper is organized as follows. In the Background, New Situation, and Problem Description sections, we present the original business environment, changes to the environment, and details of the resulting problem. In the Approach and Assumptions section, we describe how we approached the problem and the assumptions we used to solve it. In the next four sections, we present the formulation details for the three separate steps of the heuristic, as well as the overall heuristic procedure. In the two final sections, we discuss the implementation and results and follow with the challenges and lessons learned.

Background

Product Description

Each computer TMC produces begins as an order a customer places with the Dell sales organization. Sales notifies the factory, which schedules the production of the order and acquires the parts. The factory is responsible for grouping all the parts the computer requires into a kit, and assembling, packaging, and shipping the computer to the customer. Each computer requires a chassis (case) and a selection of component parts that are assembled into this chassis.

TMC groups its products into product families to aid in the production-planning process. The chassis a product requires defines the product family. Hence, the terms "chassis" and "product family" are essentially interchangeable. Each chassis has a variety of components that can be assembled with it. We will refer to a product family's chassis and set of components collectively as the product family's parts. Each family has some components that are specific to it, and some that various other families also use. Similarly, some components are limited to one family and others to several families.

Factory Configuration

The factory has multiple identical kitting lines located side-by-side. Figure 1 depicts a rough schematic of one kitting line. As the diagram shows, each kitting line has spaces designed to hold pallets of chassis and spaces designed to hold boxes of components. Each line can hold [N.sub.a] ([greater than or equal to] 1) different types of chassis (one type of chassis per pallet) and [N.sub.b] ([greater than or equal to] 1) different types of components (one type of component per box). Each chassis space on a line is called a lane, and each component space is referred to as a bin.

[FIGURE 1 OMITTED]

When the line is running, a worker selects one chassis from the [N.sub.a] lanes for the next product to be built and sends the chassis to the next worker down the line. The workers down the line select the correct components for that order from the bins and add them to the chassis. The production computer system keeps track of which chassis are in which lanes and which components are in which bins, and signals the workers to select the correct chassis and components for each order.

Setup Description

As is sometimes the case with products that can be grouped into product families (Webster and Baker 1995), kitting a product from one family immediately following another product from the same family, as in Dell's situation, does not require another setup. What is unusual about this problem is that no setup is required when kitting a product from one family immediately following a product from a different family if the chassis and parts that both families require are on the kitting line at the same time. Therefore, a setup is only required when the product to be kitted is from a different family than all the [N.sub.a] families whose parts are currently on the line.

Such a setup entails removing one family's pallet of chassis from a lane and placing the pallet of chassis required for the next order in that lane. Changing components is also necessary. Workers must remove those not needed for the families of the chassis remaining in the other lanes, or for the family being added to the line, from the bins. Then, they must place the components that the new family requires, but are not already on the line, in the bins. Finally, workers must update the production computer with the new information regarding which parts are in which spaces.

Production Planning

Initially, Dell offered customers a choice of about 10 to 12 different computer chassis, which TMC would assemble. In addition, Dell offered customers several component choices for each chassis type. With several parallel kitting lines, it was possible to place each type of chassis at the front of at least one kitting line at all times and to also place any associated components on that line. As a result, Dell could build any computer requiring any chassis and combination of components at any time without having to change any parts placed on any kitting line. The only exceptions were when the product...

View this article FREE - Now for a Limited Time, try Goliath Business News
Free for 3 Days!



More articles from Interfaces
Morehead City, NC, uses an ARIMA study to end a state moratorium on ne..., May 01, 2007
A novel problem for a vintage technique: using mixed-integer programmi..., May 01, 2007
Making decisions about safety in US ports and waterways., May 01, 2007
The US Army uses decision analysis in designing its US installation re..., May 01, 2007
Practice abstracts.(operations research methods), May 01, 2007

Looking for additional articles?
Search our database of over 3 million articles.

Looking for more in-depth information on this industry?
Search our complete database of Industry & Market reports by text, subject, publication name or publication date.

About Goliath
Whether you're looking for sales prospects, competitive information, company analysis or best practices in managing your organization, Goliath can help you meet your business needs.

Our extensive business information databases empower business professionals with both the breadth and depth of credible, authoritative information they need to support their business goals. Whether it be strategic planning, sales prospecting, company research or defining management best practices - Goliath is your leading source for accurate information.