A novel population initialization technique for Genetic Algorithm

Genetic Algorithm (GA) has been proved to be efficient at searching optimal solution among a large and complex search space in an adaptable way. The traditional GA doesn´t provide effective performance with random population seeding technique using which the population may contain poor quality individuals that takes long time to converge to an optimal solution. This motivates to devise a novel population initialization technique with the features of randomness and individual diversity. In this paper, an innovative Vari-begin and Vari-diversity (VV) population seeding technique has been proposed. Experimentation is performed on Travelling Salesman Problem instances, based on convergence rate, obtained from TSPLIB using MATLAB tool shows that the developed technique can produce the individuals with high fitness.