The multi-vehicle routing problem is to determine the optimal routes for all vehicles through a set of locations subject to vehicle capacity constraints and other restrictions. The objective is for the last vehicle to return to the depot as fast as possible. The problem is formulated as an integer programming problem with two types of variables. One is indicated by a sequence of locations, and the other by the set of the first locations arrived at on their routes. The best sequence of the former variables can be obtained by using genetic algorithms, and the latter can be optimized by applying the branch-and-bound technique.