Comparison of nongradient methods with hybrid Taguchi-based epsilon constraint method for multiobjective optimization of cylindrical fin heat sink

The primary aim of the paper is to compare the different nongradient methods of multiobjective optimization for optimizing the geometry parameters of a cylindrical fin heat sink. The methods studied for comparison are Taguchi-based grey relational analysis, ε (epsilon) constraint method and genetic algorithm. The various responses that have been studied are electromagnetic emitted radiations, thermal resistance and mass of the heat sink. Since the responses are obtained using complex simulation softwares (HFSS—Ansoft for emitted radiations and CFD—Flotherm for thermal resistance), there is no way of calculating the derivates of the objective functions. Hence, the Taguchi design of experiments design is used to derive the linear regression equations for the responses studied, which are then taken as the objective functions to be optimized. A new hybrid method known as Taguchi-based epsilon constraint method has been proposed in this paper for obtaining nondominated Pareto solution set. The results obtained using the proposed method show that the Pareto optimal set is competitive in terms of diversity of the solutions obtained. It is not likely that there exists a solution, which simultaneously minimizes all the objectives using any of the multiobjective techniques implemented. The value path analysis has been done to compare the trade-off among the design alternatives for the chosen multiple objective parameter optimization problem.