Efficient mataheuristic algorithms for the multi-stripe travelling salesman problem

Abstract

The Multi-stripe Travelling Salesman Problem (Ms-TSP) is an extension of the Tra-velling Salesman Problem (TSP). In the q-stripe TSP with  1, the objeetive function sums the costs for traveling from one vertex to each of the next q vertices along the tour. To solve medium to large-sized imstanees. a metaheuristic approach is proposed. 'The proposed method has two main components, which are construction and improvement phases. The eonstruction phase generates an initial solution using the Greedy Randomized Adaptive Search Procedure (GRASP). In contrast, the optimization phase improves it with several variants of Variable Neighborhood Search (VNS). both coupled with a technique caled Shaking Techniqune to eseape from local optima. Besides. the Adaptive Memory (AM) technique is applied to balanee between diversification and imtensification. To show the efficiency our proposed metaheuristic algorithins. we extensively implement them on benchmark instances. The results indiecate that the developed algorithms can produce efficientt and effective solutions at a reasonable computation time.