Research On Financial Escorting Vehicle Routing Problems Based On Tdvrp And Stdvrp Models

Nowadays, specialized financial escorting firms have been founded in main cities, providing service for bank branches. Although developing rapidly, financial escorting firms usually make vehicle routing decision on experience. The purpose of this research is to provide decision-making support for vehicle routing of such firms.There are three unique characteristics of financial escorting, compared to other vehicle routing problems. Firstly, it's usually done in urban network during peak hours, thus travel speed are influenced largely by departure time and random fluctuation, while time windows of are very tight. Secondly, if route failure happened, no ordinary remedy strategies can work, because each goods is specific to one customer. Lastly, vehicle routing schedule needs stability over a long horizon. The paper treats this problem as TDVRP and STDVRP models. In TDVRP model, speed is assumed a function of time to avoid passing-by problem. In STDVRP model, both time dependence and fluctuation of speed are considered.To solve the models above, the paper proposes a Multiple Population Evolution Strategy (MPES), and proves its capacity of global optimal search on Solomon's benchmark. The paper also develops test instances under time-dependent and stochastic time-dependent situations. Comparison and analysis of results prove the effectiveness of TDVRP and STDVRP models.In addition, the paper conducts a case study on real business of Vpower Financial Escorting Company in Nanshan District, Shenzhen city. The results show that the paper adds value both in theoretical innovation and practical application.