A Comparison of GA Crossover and Mutation Methods for the Traveling Salesman Problem

Abstract

The traveling salesman problem is a very popular combinatorial optimization problem in fields such as computer science, operations research, mathematics, and optimization theory. Given a list of cities and the distances between any city to another, the objective of the problem is to find the optimal permutation (tour) in the sense of minimum traveled distance when visiting each city only once before returning to the starting city. Because many real-world problems can be modelled to fit this formulation, the traveling salesman problem has applications in challenges related to planning, routing, scheduling, manufacturing, logistics, and other domains. Moreover, the traveling salesman problem serves as a benchmark problem for optimization methods and algorithms, including the genetic algorithm. In this paper, we examine various implementations of the genetic algorithm for solving two examples of the traveling salesman problem. Specifically, we compare commonly employed methods of partially-mapped crossover and order crossover with an alternative encoding scheme that allows for single point, multipoint, and uniform crossover. In addition, we examine several mutation methods, including twors mutation, centre inverse mutation, reverse sequence mutation, and partial shuffle mutation. We empirically compare the implementations in terms of the chosen crossover and mutation methods to solve two benchmark variations of the traveling salesperson problem. The experimental results show that the genetic algorithm with order crossover and the centre inverse mutation method provides the best solution for the two test cases.

Integrated Q-Learning with Firefly Algorithm for Transportation Problems

“…This led to the emergence of combined Q-learning with other meta-heuristic-based algorithms. In [9], a Q-learning-based PSO method is proposed for path planning in mobile robots. A comparative study of various meta-heuristic algorithms in TSP is conducted [10].…”

FMap: A fuzzy map for scheduling elephant flows throughjumping traveling salesman problemvariant towardsoftware‐defined networking‐baseddata center networks

“…The problem can be described as follows: Given a list of n cities (network nodes) and the distances between each pair of cities, the objective function starts finding the optimal permutation (tour) that realizes the minimum traveled distance (link weight). Furthermore, the objective function should observe the problem constraints (1) all cities should be visited; (2) no city should be visited more than once [39][40][41][42][43] ; and (3) when the destination node got visited, the tour ends, except when the problem is defined as open (in which it does not go back to the initial node).…”

Combining reinforcement learning algorithm and genetic algorithm to solve the traveling salesman problem

“…To overcome these challenges, researchers have been exploring more innovative and efficient solutions. In recent years, reinforcement learning [10][11][12] and genetic algorithms [13,14] have emerged as two distinct optimization methods that have shown excellent performance in solving combinatorial optimization problems.…”