Joint location optimization of charging stations and segments in the space-time-electricity network: An augmented Lagrangian relaxation and ADMM-based decomposition scheme

Abstract

Electric vehicles that contribute to better air quality, less noise, and low-carbon emissions are a promising selection for sustainable transportation. However, the development of electric vehicles is impeded by various factors, including driving range anxiety, long recharging period, and insufficient charging facilities. As the charging-while-driving techniques gradually mature, electric vehicles can recharge on charging stations stationarily or charging lanes (segments) dynamically. Additional charging facilities should be constructed to improve the level of charging services. Under a predefined construction budget, the difficulty is how to determine their numbers and distributions. This study jointly locates charging stations and segments by maximizing the accessibility of electric vehicles. In space-time-electricity networks, we formulate a multi-commodity network flow model with location and routing. A decomposition scheme based on augmented Lagrangian relaxation and alternating direction method of multipliers is developed to tackle this problem. The standard and augmented Lagrangian relaxed problems are decomposed into many solvable subproblems. Numerical experiments are conducted in three transportation networks, showing that the proposed method can achieve good integrality gaps.