Electric vehicle routing problem considering traffic conditions and real-time loads

Eco-driving strategies aim to reduce energy consumption (EC), greenhouse gas (GHG) emissions, and accident rates, and to promote environmental benefits. One of the efficient ways to reduce GHG emissions is using electric vehicles (EV) to replace fuel ones. Inspired by these, in this paper, we study an electric vehicle routing problem (EVRP) with the objective of minimizing the total EC of the vehicles during transportation. Specifically, we consider segmented linear speed functions that simulate traffic conditions and take into account the impact of real-time load and acceleration of EVs on energy consumption. The speed model accurately captures gradual changes in vehicle speed, reflecting real-world road conditions more realistically. We define this problem as a time and load-dependent EVRP (TLD-EVRP). This problem is difficult to solve not only because it considers both real-time vehicle speed and load conditions but also because the problem is modeled as nonlinear (with quadratic and cubic terms of vehicle speed). The goal is to minimize the EC of the electric vehicle, and this integration will improve energy efficiency and environmental benefits. To tackle this TLD-EVRP, a meta-heuristic algorithm is developed, combining a Large Neighborhood Search (LNS) with a Local Search (LS) procedure and a split algorithm for efficient route segmentation for individual EVs. The set partitioning problem (SPP) is used to recombine the routes. The goal of the algorithm is to develop a robust tool to address the complexity of TLD-EVRPs with different sizes and distributions. The proposed approach is evaluated using a small-scale real-world test case and extensive experiments are then conducted on randomly generated large-scale data. The evaluation results show that our algorithm can handle large-scale instances of up to 1000 customers and exhibits robustness on large data. By promoting energy-efficient practices in the transportation sector, the proposed methodology contributes to the achievement of green and low-carbon goals.