Description
Complex products in the manufacturing industry, such as in vehicle construction, are characterized by numerous variants and consist of a growing number of components to be procured from suppliers. In the early phase of product development it is important to solidly estimate the costs of these externally sourced components, among other things in order to evaluate development alternatives or to support negotiations with potential suppliers. The so-called cost engineering function takes over this task. However, traditional costing methods are too costly to cope with the diversity of procurement.
In order to meet this challenge new intelligent approaches are to be compiled, which increase efficiency in cost engineering in the early product development phase by (partially) automating the costing process.
In this book it is shown that a variety of methods from the field of machine learning are suitable for intelligent cost estimation and can effectively facilitate the tasks of the cost engineer. This is substantiated on the one hand by theoretical analyses of the state of the art and on the other hand by differentiated results of three practical case studies, each with numerous individual experiments.
In these case studies, solution concepts are developed and evaluated in corporate practice, which support cost engineers in cost estimations with the help of machine learning (ML) in a targeted manner. Through experimental design and a variety of test results, it is possible to carve out best practices of ML models for cost estimation of components and assemblies to be procured. Such a ML model is characterized in particular by the "best" combination of the immense variety of ML algorithms, associated training and testing concepts, feature selection methods as well as explanatory approaches of the so-called explainable artificial intelligence. In addition, valuable insights into the performance and accuracy of ML approaches are provided.
A concluding chapter shows existing challenges in the practical environment to introduce such systems for intelligent cost estimation. For this purpose, an empirical study consulting 130 managers in relevant areas is carried out. Based on the findings, the limitations and potentials of machine learning for cost engineering are presented from both a scientific and a practical perspective.
Reviews
There are no reviews yet.