Genetic algorithm-based search on the role of variables in the work hardening process of multiphase steels

Multi-objective optimizations of strength and ductility of multiphase steels are conducted using genetic algorithms (GAs), to investigate the role of the composition and process variables in their complicated work hardening process. Neural network-based computational models, describing the complex correlations between the decision parameters for processing and materials chemistry of such steels, are developed using existing data and are used for the fitness functions. The cases of both high-strength low-alloy steel (HSLA) and the transformation-induced plasticity (TRIP)-aided steel are separately studied, and the findings are compared and contrasted. The Pareto solutions are used successfully to study the role of the parameters at different combinations of strength and ductility. The findings are also utilized for qualitative assessment of the dominant mechanisms behind the work hardening of the steels.