Robust Optimization of Electric Vehicle Paths in Uncertain Environments

Abstract

In the face of uncertainty in customer demand and dynamic traffic conditions, determining the optimal logistics distribution route under deterministic conditions becomes challenging. To address this issue, we propose a novel approach by adopting Robust Optimization Models for Electric Vehicle Path Optimization. Our robust optimization model incorporates two uncertainty sets, namely the convex set and the box set, to tackle variations in demand and speed. By introducing deviation coefficients, we compare the objective function values of the robust optimization model with those of the deterministic model. This enables us to understand the trade-offs between robustness and optimality. To find solutions for various instance sizes, we apply an improved genetic algorithm to solve the constructed model efficiently. Our case study results demonstrate that while the optimal objective function value of the robust optimization model may be higher than that of the deterministic model, it ensures the feasibility of the path even amidst demand fluctuations and dynamic traffic conditions. Moreover, we analyze the economic returns of velocity and demand under both uncertainty sets with different data sizes, using the deviation coefficients. These discoveries provide valuable perspectives for pertinent departments, assisting them in rendering well-informed choices and attaining practical significance in real-world scenarios. To recapitulate, our study introduces an innovative methodology for addressing the challenges of optimizing electric vehicle routes in the face of unpredictable demand fluctuations and time-dependent speed variations. By demonstrating the effectiveness of our proposed robust optimization model, we contribute to the advancement of logistics and transportation systems in a volatile and uncertain environment.