An Efficient Solution to Travelling Salesman Problem using Genetic Algorithm with Modified Crossover Operator
Md. Sabir Hossain,Ahsan Sadee Tanim,Sadman Sakib Choudhury,S. M. Afif Ibne Hayat,Muhammad Nomani Kabir,Mohammad Mainul Islam
DOI: https://doi.org/10.24003/emitter.v7i2.380
2019-12-01
EMITTER International Journal of Engineering Technology
Abstract:The traveling salesman problem (TSP) is a famous NP-hard problem in the area of combinatorial optimization. It is utilized to locate the shortest possible route that visits every city precisely once and comes back to the beginning point from a given set of cities and distance. This paper proposes an efficient and effective solution for solving such a query. A modified crossover method using Minimal Weight Variable, Order Selection Crossover operator, a modified mutation using local optimization and a modified selection method using KMST is proposed. The crossover operator (MWVOSX) chooses a particular order from multiple orders which have the minimum cost and takes the remaining from the other parent in backward and forward order. Then it creates two new offspring. Further, it selects the least weight new offspring from those two offspring. The efficiency of the proposed algorithm is compared to the classical genetic algorithm. Comparisons show that our proposed algorithm provides much efficient results than the existing classical genetic algorithm.