Decision support system in the make to order manufacturing system

Abstract

Current methods for estimating the cost and time are based on decomposing the product into elements, followed by cost estimation of each element and summing of the other costs. As element, we can consider a product component, a manufacturing process component or an activity component. To estimate the cost for each element, its features that are closely related to cost or time are used. With a few exceptions, the estimation methods lead to estimation without a mathematic model describing the relation between cost or time and the element’s features. Moreover, these methods have a slight adaptation capacity to different specific situations because the information that is provided in order to make estimations is general and does not adapt to a specific case. Therefore, in this paper, the cost and time will be estimated by a set of appropriate techniques which are based on neural modeling and k-nearest neighbor regression. Each of these techniques cover, a range of specific cases.