Solving the consistent Vehicle Routing Problem with Profit using Simulated Annealing

Abstract

Consistent VRP with profits is a variant of vehicle routing problems, it was introduced in 2019. In this problem, each instance is defined by a depot, a group of customers (frequent customers and non-frequent customers) and a group of vehicles with a specific capacity. In this variant, each customer has a demand, a service time, and a fixed profit. The goal is to find the best set of paths for the group of vehicles, while respecting the constraints of consistency, which are Person consistency, Time consistency, and consistency in the delivered quantity. In this work, we used the simulated annealing meta-heuristic for solving the consistent VRP with profit. In the definition of the neighborhood, we have implemented several operators noted 2-Opt move, 1-1 exchange move and 1-0 exchange move. In the computational experiments, we have compared the results obtained by the simulated annealing algorithm with those of an Adaptive Tabu Search and also with those of branch and-bound exact method. The comparison showed that SA performs good results for small-sized instances. For the large-sized instances we found that SA is not enough efficient to solve this NP-complete problem, comparing with ATS.