Research On Hub-and-spoke Network Planning For Road Express

In recent years, with the great development and progress of our country's road express freight business, an efficient and effective, wholesome and perfect transportation network attracts more and more attention from the road express freight enterprises.Road express freight enterprises which mainly handle less-than-truck load (LTL), usually adopt hub-and-spoke network to minimize operation cost by substituting high transportation cost with lesser transferring cost. At present, the researches on hub-and-spoke network mainly concern over the hub location and network load planning, and some researchers recently begin to study the combined planning, but these researches emphasize particularly on pure hub-and-spoke without stopovers, which does not accord with feature of LTL. Moreover, the load planning is a multi-decision problem, various decision-makings have relativities, but preceding researches mostly solve problem by phases, the result is not globally optimal solution in principle.This paper carries out researches on hub-and-spoke network with multi-hub and stopovers, and divides the network planning into two stages: hub location and route optimization. Firstly, factors influencing network design are extracted, such as freight volume, geographic location and service timeliness. A new method combining AHP (Analytic Hierarchy Process) with FCM (Fuzzy Clustering Method) is proposed to confirm count and locations of hubs. Secondly, aiming to minimize the operation cost, the paper extracts related cost, and establishes the vehicle routing model with time constraint and capacity constraint according to in-depth study of freight flow. The assignment relationships between spokes and hubs, the trunk count of various type and routes are decision variables. Lastly, an algorithm is designed to obtain the best network design program.In the process of model solving, a two-double GA (Genetic Algorithm) is designed to solve the problem. Algorithm synthetically considers the assignment relationships between the spokes and hubs and vehicle routing optimization, which the outer GA optimizes the assignment relationships, the inner GA optimizes vehicle routing. Type-choosing relies on the selection strategy based on roulette of GA. At last, based on the principle of minimum cost, the paper determines the best assignment program and the optimal vehicle routing accordingly. The paper takes an express freight enterprise for example, and verifies the validity of this program. From the result of calculation, it can be seen that the new network after optimizing can save cost by 6.30% than before. After adjusting, the cost can be saved by 9.17% than before. The application value of the network design program and practicability of the model presented are validated through this case study.