Performance Comparison of Different Operation Techniques in Genetic Algorithm towards Benchmark Functions

Abstract

Optimisation becomes a popular topic in various types of research due to its functionality and characteristics to optimise the result based on the resources available. The importance of using optimisation in solving the case studies has been revealed by researchers in the past. In addition, Genetic Algorithm (GA) is also one of the famous methods that can be used in solving the optimisation problem. However, the performance of GA is always not considered by the researchers due to the limited of knowledge on this method. Therefore, this research aims to analyse and identify the best combination of operation techniques in GA by evaluating their performance using average fitness values. 3 selection, crossover, mutation, and 1 replacement operation techniques are applied and tested using 10 benchmark functions. The experimental results show that GA25 is the best GA model compared to the other models because it is able to obtain 8 out of 10 minimum average fitness values. Therefore, GA25 with Tournament Selection, Uniform Crossover, Flipping Mutation and Weak Parent Replacement is considered as the best combination operation techniques model. The output of this research is able to assist the researchers to gain a better understanding of GA while implementing the method in their research.