Cost Estimation in a Capacitated Environment

Introduction

Nowadays, in the global competitive market, companies' prosperity is strongly dependent on their ability to accurately estimate product costs. This is especially true for firms operating in a make-to-order environment. For such firms, a small error in a price quote, resulting from erroneous product cost estimation, may make the difference between being awarded and losing a contract.

During the last decade, the relative weight of direct labor costs in manufacturing has dramatically diminished, while the relative weight of indirect costs has increased (Gunasekaran, Marri & Yusuf, 1999). Therefore, allocating indirect costs to products, without taking into account shop floor capacity, may lead to erroneous estimations. Despite this, most existing cost estimation models assume unlimited shop floor capacity. There are many such models in the literature. These models use information about the products, the materials and the production processes. Common approaches are:

• •

  Parametric cost estimation models that are based on:

  • o

    Regression analysis (Cochran, 1976a, b; Ross, 2002).

  • o

    Fuzzy logic (Jahan-Shahi, Shayan & Masood, 2001; Mason & Kahn, 1997).

  • o

    Minimization of Euclidean distance between the estimated cost and its actual value (Dean, 1989).

  • o

    Neural networks (Bode, 2000; Lin & Chang, 2002; Shtub & Versano, 1999; Smith & Mason, 1997).

• •

  Bottom-up cost estimation models, in which the total cost is the sum of detailed components (Rad & Cioffi, 2004; Son, 1991; Stewart, 1982).

• •

  Group technology cost estimation models that use the similarity between products from the same family (Geiger & Dilts, 1996; Jung, 2002; Ten Brinke, Lutters, Streppel & Kals, 2000).

• •

  Hybrid cost estimation models that combine some of the models described above (Ben-Arieh, 2000; Sonmez, 2004).

Parametric, bottom-up, group technology and hybrid cost estimation models use only information about the product, the materials from which it is made, and the production processes required for its manufacture. None of the above cost estimation methods takes considers the available capacity on the shop floor. The assumption is that the available capacity is sufficient. However, in reality, one must deal with finite capacity and dynamic workloads, which may change over time.

Here we assumed that the total cost of the product is a function of the load on the shop floor (which is made up of the orders waiting to be manufactured or actually being manufactured in a certain time period). Specifically, we assumed that the cost of producing an order when the load is high is different from the cost of the same order when the load is low and most resources are idle. It is our contention that ignoring the load on the available capacity distorts the product cost estimation and may lead to wrong decision-making.