In this paper, we address the problem of evaluating the benefits of considering the multiple use of a set of homogenous vehicles in the routing problem. Based on a mixed-integer linear programming model, a Tabu Search (TS) algorithm was presented for the multiple-trip vehicle routing problem. Further discussion shows that the multiple use of the vehicles is not always cost effective. Two criteria are given to determine, under the given situation, whether or not to take the multiple use of the vehicles into consideration. Our conclusions are verified by the computational results of a series of test problems. It is shown the algorithm produces high quality solutions, and the criteria can save the computational times to a certain extent. 

Key words: vehicle routing problem; Tabu search algorithm; multiple trips.