Computational Experiments on Metaheuristic Techniques for Combinatorial Problems

Abstract

This thesis investigates the application of metaheuristic techniques in addressing single and multi-objective combinatorial optimization problems, focusing on the essential role of efficient solution strategies. Firstly, this thesis underscores the importance of meticulously designed solution frameworks through an in-depth analysis of three specific combinatorial optimization problems: production planning of a petrochemical industry, scheduling jobs on dissimilar parallel machines, and optimization of a compression-absorption cascaded refrigeration system. The thesis also addresses the limitations of metaheuristic techniques in solving computationally expensive problems by proposing variants that leverage parallel computing architectures. Additionally, this work emphasizes the necessity of precise implementations of metaheuristic techniques to ensure optimal effectiveness in their application.