To attain the goal of cost minimization, a vehicle routing model for cargo transport O2O platforms were established in this study. In consideration of differences in vehicle origin in the traditional vehicle routing problem as well as the one-to-one corresponding relationship between cargo owners' pickup and delivery points, constraint conditions such as half-open, multiple depot, multiple vehicle type, origindestination pair, loading limit, and soft time window constraints were introduced into the proposed model. Given the model characteristics, an improved genetic algorithm, which is commonly used in vehicle routing problem, was used as the solving tool. The nearest matching method currently used by cargo transport O2O platforms was simulated using the simulation software AnyLogic. Moreover, vehicle-cargo orders on a platform within a certain time period were selected and allocated, and a matching scheme was obtained. Then, the optimized matching scheme for the same order was calculated using the improved genetic algorithm. Results show the comprehensive cost obtained by the improved genetic algorithm is 21.14 % lower than that of obtained by the nearest matching method.

No abstract