Cost-based Procedure for Multi-response Parameter Design Problems Using GEP, Taguchi Quality Loss, and PSO: Case Study on Heat Sink Design

Abstract

Most modern products/processes usually have several quality characteristics that must be optimized simultaneously; this is called a multi-response parameter design problem. To overcome shortcomings in the literature, including insufficient accuracy of second-order polynomials, subjective determination of relative weights and shape coefficients, and non-consideration of manufacturing or material costs, this paper proposes a cost-based procedure for resolving multi-response parameter design problems using gene expression programming (GEP), Taguchi quality loss, and particle swarm optimization (PSO). A case study with the aim of optimizing the design of a heat sink applied to a high-power MR16 LED lamp was used to demonstrate the proposed procedure. The experimental results indicated that the proposed procedure can provide highly robust settings for design parameters that can maximize the thermal performance and minimize the actual material cost of a heat sink. Furthermore, decisionmakers no longer need to subjectively determine the relative weight of each response. Therefore, the proposed approach can be considered to be feasible and effective; it has the potential to be a useful tool for resolving general multi-response parameter design problems in the real world.

Hybridizing a fuzzy multi-response Taguchi optimization algorithm with artificial neural networks to solve standard ready-mixed concrete optimization problems

“…For determining the optimal factor levels on concrete quality, experimental design methodologies such as Response Surface Methodology (RSM), Taguchi Methods and factorial design applications are widely used in literature [4][5][6][7][8][9][10][11][12][13][14]. Multi-Attribute Decision Making (MADM) methods are also combined with Taguchi methods where multiple responses are involved [15][16][17][18].…”