O.R. Applications
Analysis of an industrial component commonality problem

https://doi.org/10.1016/j.ejor.2007.01.008Get rights and content

Abstract

We discuss a case study of an industrial production-marketing coordination problem involving component commonality. For the product line considered, the strategic goal of the company is to move from the current low volume market to a high volume market. The marketing department believes that this can be achieved by substantially lowering the end products’ prices. However, this requires a product redesign to lower production costs in order to maintain profit margins. The redesign decision involves grouping end products into families. All products within one family use the same version of some components. This paper fits in the stream of recent literature on component commonality where the focus has shifted from inventory cost savings to production and development cost savings. Further, we consider both costs and revenues, leading to a profit maximization approach. The price elasticity of demand determines the relationship between the price level and number of units sold. Consequently, we integrate information from different functional areas such as production, marketing and accounting. We formulate the problem as a net-present-value investment decision. We propose a mixed integer nonlinear optimization model to find the optimal commonality decision. The recommendation based on our analysis has been implemented in the company. In addition, the application allows us to experimentally validate some claims made in the literature and obtain managerial insights into the trade-offs.

Introduction

We study a real life component commonality problem for a large international manufacturing company, a major player in the industrial power tools industry. This study relates to one specific machine type and product line, called the T-range. The machines range from small units for domestic use to very large models for industrial applications. The kilowatt (kW) size is the main indicator of the power and size of the machine. The company launched a project to develop a completely new and cheaper design for the T-range. The goal of the project was to boost sales volumes and profits of the product line while at the same time increasing the quality and benefits for the customer. This can be achieved by lowering market prices which requires a decrease of production costs to maintain profit margin. The redesign will lead to the required reduction in total manufacturing costs.

Currently, this T-range contains eight models, which we label T1–T8 with increasing kW size. The T-range is split up into families, which are defined as a group of end products sharing the same component. There are three families: (1) T1–T4, (2) T5 and T6, and (3) T7 and T8. The simplified Bill of Material (BOM) for the current range is given in Fig. 1. The frame (F) is the housing and all mechanical and electrical appliances that form a machine. The stage (S) is an essential component for the product. The frame and stage are common components for each family and are always dimensioned to fit the requirements of the machine with the highest performance in the family. For example, all the models in the first family use the same stage (S4) and frame (F4) which are specifically dimensioned for T4, the highest kW machine in that family. The engine (E) is specific for each model. The redesign project includes a review of the actual range split of the families and a proposal for a new split. The component commonality decision must therefore determine which end products will use the same stage and frame in the redesigned product line.

The main elements and relationships of this problem are visualized in Fig. 2. The details will be explained in Section 2. The key cost factors that drive the commonality decision are the unit production costs (a) and the development costs (b). The component commonality decision has opposite effects on these two cost factors, resulting in a trade-off. It is the company’s practice to determine the price (d) by adding a fixed mark-up (c) to the costs (a). However, cost minimization alone is not relevant here. The number of units sold (f) is not fixed, but depends on the price (d) through the price elasticity (e). Taking both costs and revenues into account leads to a profit maximization approach and therefore the design and pricing issues must be considered simultaneously. We formulate this problem as an investment decision. The optimization of the component commonality decision weighs the investment costs, i.e. the development costs for the common components (b), against the future extra profit (g) to calculate the net present value (h) of the project.

Our paper has several contributions. First, we present an industrial component commonality problem with several new aspects. Our real life case supports the recent shift in focus in the literature on component commonality from the savings in inventory costs to the impact on production and development costs, as these costs have a much bigger impact. Furthermore, cost and revenue sides are integrated, leading to a profit maximization approach. The price sensitivity provides the link between the price level and sales. This component commonality problem requires a cross-functional approach, integrating marketing, production and accounting information within an investment analysis framework. Further, our industrial problem considers two different components simultaneously. Second, we develop a mathematical programming model for this problem. This formulation is flexible, allowing us to model a variety of industrial constraints. In an extension of our model, we explore the problem of product line pricing by treating the selling price as a decision variable. Third, our industrial case allows us to experimentally validate earlier claims made in the literature and we also perform a series of experiments to obtain general insights into the problem.

The paper is organized as follows. In Section 2, we give a detailed analysis of the case problem and relate it to the relevant literature. All the factors mentioned in Fig. 2 will be discussed in detail here. Section 3 describes the general mathematical formulation. In Section 4, we extend the model with additional industrial constraints. Finally, results and managerial insights for our specific case are discussed in Sections 5 Implementation and results, 6 Managerial insights.

Section snippets

Cross-functional analysis

The design of families and common components is not a stand-alone decision. It affects the entire value chain of the company. Eynan and Rosenblatt (1996) remark: “The decision on whether to employ component commonality is usually made at the design stage. In general, the component commonality decision affects, and is part of, other major functions of the business such as purchasing, production and marketing.” In order to make a good decision, information from different departments is needed.

An initial model

Essentially, the product range split problem as described here is an assortment problem. In such a problem we have a set of sizes of a product with associated demands. For practical reasons, we cannot make all the different sizes and we have to choose a subset of sizes which will be produced. Demand for a size which is not produced can be met by a larger size of the chosen subset. Such a substitution involves a substitution cost (Pentico, 1976). In our case we have such an assortment problem at

Further industrial considerations

During the study many specific industrial considerations emerged and we had to extend our basic model to include all real life issues. A good choice of the variables in the basic model made it possible to easily implement new constraints. We will discuss these industrial considerations without too much detail. We have incorporated the following additional engineering and marketing issues into our model:

  • We consider two common component types simultaneously, namely the stage and frame. As a

Implementation and results

This research is an extension of previous studies on the use of common components done by the consultancy firm OM Partners (Belgium) at other subdivisions of the company in 1990 and 1994. This is the third time that a similar problem was studied for the company, and after this project, a new study for a different product line was undertaken. This illustrates the relevance and value added of the study and indicates that this is a generic problem for the manufacturer. The new critical element in

Managerial insights

In the product range split problem, we must find the optimal split of a product range into families, where all the models in the same family share the same version of a component. A similar problem has also been studied by Fisher et al. (1999). They state that this as a relevant problem for products with the following properties:

  • (1)

    the models in the range make use of the same type of component,

  • (2)

    the models can be ranked by performance requirement of the common component,

  • (3)

    the common components are

Conclusion

In this paper, we describe a real life product range split problem, which is essentially an assortment problem and a component commonality problem. We propose a mixed integer nonlinear optimization model for this problem. The component commonality decision affects both the production costs and development costs. The key feature of our analysis is the incorporation of the demand relationship, where the volume sold is price dependent. Integration of both costs and revenues is necessary for the

Acknowledgements

We thank one referee for providing many suggestions to improve the exposition of the paper and frame it better as a case study.

References (26)

  • A. Dogramaci

    Design of common components considering implications of inventory cost and forecasting

    AIIE Transactions

    (1979)
  • A. Eynan et al.

    Component commonality effects on inventory costs

    IIE Transactions

    (1996)
  • M. Fisher et al.

    Component sharing in the management of product variety: A study of automotive braking systems

    Management Science

    (1999)
  • Cited by (0)

    View full text