A Comparative Study on using Genetic Algorithm with DEAP framework on Different Optimization Problems

Evolutionary Algorithms (EAs) are the solution approaches for solving optimization problems using the classical generate-and-test method. These algorithms are widely used by the practitioners for solving the optimization problems and are studied by the researchers for their advancements. Few highlighted advancements in the domain of EAs are distributed EAs, ensemble EAs, modular EAs, and tuning free EAs. Nevertheless, for the practitioners to get the clarity on the working principle of EAs, the literature is lacking with a comprehensive article reporting their performance on different problems. The research study presented in this paper is an attempt to alleviate this shortcoming. This paper aimed at presenting in detail the working principle of Genetic Algorithm (GA) on different problems which are generally used as reference problems for design and analysis of algorithms. The problems considered in the study are one-max problem, 0/1 Knapsack problem and N Queen problem. The experimental setups are carefully designed, for these three problems, and are solved by implementing GAs on the DEAP (Distributed Evolutionary Algorithms in Python) framework. The performance of the GAs is studied based on the quality and speed performance metrics and the inferences are presented in this paper.