Modeling of Gelcast Ceramics using GP and Multi Objective Optimization using NSGA-II

Abstract This proposed work introduces a novel integrated evolutionary approach and its applications for modeling and optimization of important manufacturing process namely gelcasting. Genetic programming (GP) is an evolutionary algorithm which uses principle similar to Genetic algorithms (GA) to model highly non-linear and complex processes resulting in accurate and reliable models. For developing models, GP method makes use of experimental data generated from the process. For gelcasting process input variables are solid loading, monomer content and ratio of monomers and performance measures are flexural strength and porosity. As the chosen performance measures are opposite in nature, there cannot be a single optimization solution. Hence the problem under consideration is to be formulated as multi objective optimization problem and solved using NSGA-II algorithm to retrieve the Pareto optimal front. Pareto set of process parameters in a gelcasting process in multi objective optimization of flexural strength porosity are obtained by executing these novel algorithms in a single run.