A hybrid large neighborhood search for the static multi-vehicle bike-repositioning problem

•This paper proposes a hybrid large neighborhood search to solve the multi-vehicle static repositioning problem.•Several removal and insertion operators are proposed to diversify and intensify the search.•A simple tabu search is further applied to the most promising solutions.•Results show that the heuristic outperforms both the CPLEX and the math heuristic by Forma et al. (2015) [Transportation Research Part B 71: 230–247].•The average improvement over the math heuristic is 1.06% using only a small fraction of computing time. This paper addresses the multi-vehicle bike-repositioning problem, a pick-up and delivery vehicle routing problem that arises in connection with bike-sharing systems. Bike-sharing is a green transportation mode that makes it possible for people to use shared bikes for travel. Bikes are retrieved and parked at any of the stations within the bike-sharing network. One major challenge is that the demand for and supply of bikes are not always matched. Hence, vehicles are used to pick up bikes from surplus stations and transport them to deficit stations to satisfy a particular service level. This operation is called a bike-repositioning problem. In this paper, we propose a hybrid large neighborhood search for solving the problem. Several removal and insertion operators are proposed to diversify and intensify the search. A simple tabu search is further applied to the most promising solutions. The heuristic is evaluated on three sets of instances with up to 518 stations and five vehicles. The results of computational experiments indicate that the heuristic outperforms both CPLEX and the math heuristic proposed by Forma et al. (2015) [Transportation Research Part B 71: 230–247]. The average improvement of our heuristic over the math heuristic is 1.06%, and it requires only a small fraction of the computation time.