Dynamic Vehicle Routing Solution in the Framework of Nature-Inspired Algorithms

Abstract

Vehicle Routing Problem (VRP), a well-known combinatorial optimization problem had been presented by Dantzing and Hamser in 1959. The problem has taken its inspiration from the transport field. In real field environment, a lot of variants of the problem exist that actually belongs to the class of NP-hard problem. Dynamic Vehicle routing problem (DVRP) is one of the variant of VRP that varies with respect to time. In DVRP, new customer orders appear over time and new route must be reconfigured at any instantaneous time. Although, some exact algorithms such as dynamic programming methods, branch and bound etc. can be applied to find the optimal route of a smaller size VRP. But, These Algorithms fail to give the solution of existed model of VRP in real field environment under given real time constraints. Courier services, dial a ride services and express mail delivery etc. are the few examples of real field environment problems that can be formulated in the form of DVRP. In this chapter, A novel variants of DVRP named as DVRP with geographic ranking (DVRP-GR) has been proposed. In DVRP-GR, geographical ranking, customer ranking, service time, expected reachability time, customer satisfaction level have been optimized. A solution of DVRP-GR using seed based particle swarm optimization (S-DVRS-PSO) has been also proposed. The simulations have been performed using customized simulator developed in C++ environment. The data sets used in the simulations are OMK-01, OMK-02 and OMK-03 generated in real vehicular environment. The solution of the proposed algorithm has been compared with the randomized solution technique. Analysis of the simulation results confirms the effectiveness of the proposed solution in terms of various parameters considered viz. number of vehicles, expected reachability time, profit and customer satisfaction.