Integrating the Differences among Drivers in Determining the Optimal Path for Green Vehicle Routing
Al Hla, Yasmin
An-Najah National University
Vehicle Routing Problem (VRP) is one of the most common real-world operations research applications that grasped a rich attention from researchers in order to develop as much realistic models as possible. Although researches have been conducted to solve different variants of VRP model, richer models are still required to simulate more real-life circumstances. For instance, the emerging unexpected conditions (i.e.: accidents, or unexpected congestion); and dealing with over-capacitated acquired fleet of trucks. More importantly, VRP models have been proposed as isolated from the most effective factor on the success of VRP plan on ground; who is the driver. Therefore, this research spots the light on the effect of driver behavior on the optimal VRP plan, and evidences are given by figures to convince decision makers with the possibility of integrating such factor provided with the significance of the accompanying effects. The level of autonomy of making logistical decisions such as speed or route changing decisions for both planner and drivers have been represented in the model by involving risk taking parameters, and the effect of changing the level of autonomy on VRP total costs has been investigated using a sensitivity analysis. Also, in order to enhance the model configurations’ practicability; the idea of “ride sharing” is introduced by involving not only full time regular drivers to serve, but also occasional drivers set to be available to serve when shortages happens in logistical services, or when remote orders are received from rural or country-side areas as uncommon destinations for regular drivers. For the purpose of ensuring environmental friendly logistical practices, the policy of velocity maximization has been used as well as imposing environmental penalties on the chosen rate of velocity associated with certain fuel consumption rate. Whereas the proposed VRP model satisfies both the driver by assigning a certain level of autonomy; and the firm’s financial objectives via total costs minimization; it additionally accounts for the consumed energy during serving customers in order to optimize the service time. A numerical instance with a hypothetical data set has been solved by Eclipse Java 2018-9 solver by using two heuristic methods which are adaptive solving algorithms and are able to find a local optimal solution (i.e.: the Greedy, and the Intra-route neighborhood heuristic), both revealed the same near-optimal solutions. Such VRP modeling and results have been used as a proof of concept to verify the proposed VRP model. Ultimately, the results are analyzed sensitively and show that the resulted insignificant increase in VRP costs due to assigning different levels of autonomy for drivers are still reasonable, as the total costs’ objective function weight has a mere effect on the total optimal solution, while that for the energy consumption function has the largest effect.
Integrating the Differences among Drivers in Determining the Optimal Path for Green Vehicle Routing Problem