Genetic algorithm based on enhanced selection and log-scaled mutation technique

Abstract

In this paper, we introduce the selection and mutation schemes to enhance the computational power of Genetic Algorithm (GA) for global optimization of multi-modal problems. Proposed operators make the GA an efficient optimizer in comparison of other variants of GA with improved precision, consistency and diversity. Due to the presented selection and mutation schemes improved GA, as named Enhanced Selection and Log-scaled Mutation GA (ESALOGA), selects the best chromosomes from a pool of parents and children after crossover. Indeed, the proposed GA algorithm is adaptive due to the log-scaled mutation scheme, which corresponds to the fitness of current population at each stage of its execution. Our proposal is further supported via the simulation and comparative analysis with standard GA (SGA) and other variants of GA for a class of multi-variable objective functions. Additionally, comparative results with other optimizers such as Probabilistic Bee Algorithm (PBA), Invasive Weed Optimizer (IWO), and Shuffled Frog Leap Algorithm (SFLA) are presented on higher number of variables to show the effectiveness of ESALOGA.