Machine Learning goes Measure Management: Leveraging Anomaly Detection and Parts Search to Improve Product-Cost Optimization

Machine Learning goes Measure Management: Leveraging Anomaly Detection and Parts Search to Improve Product-Cost Optimization

In many industries, particularly discrete manufacturing, companies can benefit from conducting product-cost optimization as early as possible. Given the amount of data to be analyzed in the costing process, the lack of dedicated information system support, and the pressure to quickly estimate the cost of new products, the potential for cost optimization is often underexploited. In this paper, we present an approach for leveraging machine learning capabilities, including similarity and anomaly analysis, to improve the identification of product-cost optimization potentials and therefore, improve the quality of early cost estimates. For the approach to succeed, however, ongoing training of a model based on a high-quality dataset is crucial. Thus, we also propose the machine learning approach's integration with our long-term research project toward improving the management of cost optimization during product development.