Joint optimization of quality and cost in brass casting using stochastic programming. Issue 10 (2nd October 2020)

• Record Type: Journal Article
• Authors: Sakalli, Ümit Sami; Birgören, Burak
• Is Part Of: Engineering optimization. Volume 52:Issue 10(2020)
• Page Start: 1645
• Page End: 1657
• Publication Date: 2020-10-02
• Subjects: Joint optimization -- stochastic programming -- process capability -- brass casting -- blending problem; Engineering design -- Periodicals; Mathematical optimization -- Periodicals; 620.0042
• DOI: 10.1080/0305215X.2019.1668933
• Languages: English
• ISSNs: 0305-215X

Abstract

Abstract : A critical process in brass casting is the blending of pure and scrap materials to satisfy specified metal ratios. The primary focus in such blending problems has always been cost minimization. The optimal blends produced by mathematical models use large amounts of scrap materials, which are cheaper but have high variations in ingredient ratios. This gives rise to quality problems. This study aims at joint optimization of cost and quality. A chance-constrained nonlinear mathematical model is developed for maximizing the minimum process capability level for a fixed cost. Then parametric programming is used to run the model for different costs to produce a Pareto-optimal frontier. An application to data from a brass factory showed that the frontier is highly nonlinear, enabling the decision maker to select a competitive process capability and cost value combination. The proposed approach is applicable to any blending problem in which ingredient amounts have statistical variation.