Scheduled service network design of long-haul freight transportation based on the segment subcontract

Abstract

A new service network design problem is introduced in this work. Compared to the traditional service network design problem, the synchronization departure time for commodity and vehicle at an arc is unknown in advance in our work. The transportation cost is charged based on the number and type of vehicles, rather than on their fixed and variable cost. Vehicles serving single-directional arcs are not obligated to return to its start node. The rent cost of a vehicle is given a discount if there are commodities assigned to its backhaul trip. To achieve the objective of optimizing the distributed paths of commodity and vehicle scheduling, a mathematical model based on the path with a minimum total cost is formulated. Optimizing commodity departure time on each arc and assigning the vehicles are essential for addressing the problem. Aimed at them, a recursion method is proposed to tackle the synchronization departure time for commodity and vehicle at each arc, and a heuristic based on the decision tree is proposed to solve the vehicle assignment problem. The effectiveness of the proposed approach is assessed through its comparison with the branch-and-bound approach of CPLEX 12.5. It is found that vehicle backhaul discount, commodity transshipment and vehicle types have a great impact on cost by the computational result of a real case.