.png){kind=icon}
.png){kind=icon}
Return
Gene Expression Programming-Based Ride Insert Policy for Online Electric Vehicle Ride-Hailing Optimization
.png){kind=icon}
PRE
AI
DOI：10.1109/TITS.2025.3618704.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
With the popularization of electric vehicles (EVs), the electric vehicle routing problem (EVRP) has gained significant attention. Integrating EVs into ride-hailing services has emerged as a promising trend, but few studies on EVRP comprehensively address key EV ride-hailing features. These features include the constraints of vehicles being unoccupied during charging, the dynamic nature of ride requests, ride-sharing constraints, and so on. Moreover, traditional EVRP algorithms face challenges related to high computational complexity and the need for real-time responsiveness in large-scale and dynamic environments. To address these gaps, this paper introduces the electric vehicle ride-hailing model (EVRHM), a variant of the EVRP designed to incorporate relevant constraints and reformulate the objective function to accommodate EV ride-hailing services better. Subsequently, we propose an online optimization algorithm for the EVRHM based on ride insert policies (RIPs) evolved by genetic programming hyper-heuristic (GPHH). Specifically, a low-level heuristic template is developed to insert dynamic ride requests. When a new ride request is received, the template employs the proposed RIP to insert the ride into the confirmed routes, considering both the current environmental state and ride-specific information. This approach eliminates the need for time-consuming re-optimization processes, enabling real-time responses in dynamic environments. Furthermore, gene expression programming (GEP) within the GPHH framework is designed to learn RIPs, reducing manual effort while improving both the generalization ability and performance of the RIPs. The experimental results demonstrate that the proposed algorithm is effective and efficient in solving EVRHM, with the evolved RIPs exhibiting superior generalization ability.
Keywords:
Electric vehicles and electric mobility
genetic programming
hyper-heuristic
online
Journal
IF:
8.4
Papers:
9.5K
Citations:
6.3W