Introduction to Computational Optimization Models for Production Planning in a Supply Chain

Frontmatter

1. Introduction

Abstract

Supply chain management rose to prominence as a major management issue in recent years. While the focus of managing a supply chain has undergone a drastic change as a result of improved information technology, production planning remains a critical issue. The ability to instantaneously exchange information along with increased computational power has enabled the use of sophisticated optimization software.

Stefan Voß, David L. Woodruff

2. Optimization Modeling

Abstract

A model captures the essential features of something without actually being the thing itself. Some models capture the shape and proportion of a physical object, but at a different scale and without the functionality. An example is a plastic model of a jet airliner. Some of these models are used as toys, but others are used to study air flows using a wind tunnel. When creating the model, some details have to be carefully reconstructed and others can be ignored entirely.

Stefan Voß, David L. Woodruff

3. Starting with an mrp Model

Abstract

Rather than creating a model from scratch, we begin with a venerable model called materials requirements planning. The model is often referred to as mrp with lower case letters (or sometimes “little-mrp”) to make clear the distinction between mrp and MRP II. We will look at MRP II later.

Stefan Voß, David L. Woodruff

4. Extending to an MRP II Model

Abstract

MRP II was inspired by shortcomings in mrp, and as a result the data processing orientation is preserved in MRP II. As was the case with mrp, we first explain the concepts behind MRP II, then we develop an optimization model to mimic and improve its behavior. After we have this model in hand, we extend it to produce a model that can give us production plans that trade off alternative capacity uses, holding inventory and tardiness in an optimized way. The letters MRP in MRP II stand for Manufacturing Resources Planning to make it clear that resources are considered in addition to materials as in mrp. The word “resource” is used to emphasize that any type of productive capability can be considered, not just machines. The Roman number II is intended to make it clear that it is an extension to materials requirements planning (mrp).

Stefan Voß, David L. Woodruff

5. A Better Model

Abstract

The data processing approach to MRP II is to use a straightforward algorithm to create an mrp solution and then to see if the plan is feasible with respect to capacity. In keeping with a data processing or information technology mentality, the users can also be given a lot of information to help them change the input data. The optimization approach that we are developing here asks for information about objectives and tries to obtain a good or optimal plan automatically.

Stefan Voß, David L. Woodruff

6. Extensions to the Model

Abstract

In the previous chapter we extended well beyond the data processing concepts embodied by MRP II to make use of the capabilities of an optimization model. In this chapter, we continue this course by describing a number of important extensions to the model. Depending on the planning environment, some or all of these extensions may be needed to produce a useful model.

Stefan Voß, David L. Woodruff

7. Implementation Examples

Abstract

The models developed in this book can be translated more or less directly into computer languages that have been developed for optimization modeling. We provide implementations of the first three models, mrp, MRPII and SCPc, using some popular modeling languages: AMPL, GAMS, MPL, OPL, and Mosel. Implementations of additional models and information about additional modeling languages are available on the authors’ web site, which is http://faculty.gsm.ucdavis.edu/^~dlw/scm.html.

Stefan Voß, David L. Woodruff

8. Solutions

Abstract

We have now developed models with reasonable detail to be used for supply chain planning. Once we have a model, we need to get the data for it, find solutions to it, and perhaps provide information about the solutions to the software or people responsible for detailed planning and scheduling. The data must either come from ERP systems that store performance data or the data must be estimated by production and engineering staff. If a solution is to be implemented, then it must be provided to the ERP system or to production schedulers.

Stefan Voß, David L. Woodruff

9. Some Stochastic Extensions

Abstract

Now that we have developed some practical models for production planning within a supply chain and have outlined solution methods, we pursue some topics that are very uncertain. The intention of this chapter is to address some issues that are not included in commercial supply chain planning software and are just barely being addressed in the research literature. As such, they are speculative research topics. The models and solution methods that we describe here may, or may not, ultimately be adopted. However, the modeling issues are critical and must be addressed.

Stefan Voß, David L. Woodruff

10. Research Directions and References

Abstract

The material that we have provided in the previous chapters has its foundations in existing literature, which will be referenced and put into perspective in this chapter. This also includes some historical remarks which should allow the reader to follow the evolution of supply chain planning up to today.

Stefan Voß, David L. Woodruff

Backmatter