Scholarlink

<p>Electric car-sharing systems (ECS) are increasingly adopting real-time pricing (RTP) to better respond to fluctuating demand and diversify service offerings for enhanced competitiveness. This study evaluates the performance of self-service and door-to-door ECS service strategies under RTP. To capture the complexities of real-world operations, we first construct realistic ECS networks by integrating point-of-interest data and graph neural network training. A duopoly game model is then developed, where two types of firms adopt distinct service strategies and adjust prices in real-time to maximize profits. Using deep reinforcement learning, the firms’ decisions are simulated and optimized across three ECS networks—Chongqing, Wuhan, and a synthetic network. Performance is evaluated across four key metrics: pricing, consumer utility, firm profit, and market share. The results reveal that door-to-door services outperform self-service models in terms of market share and profitability, as consumers prioritize convenience over price. The findings provide strategic insights for ECS operators.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Performance of heterogeneous service strategies for a real-time pricing electric car-sharing system

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2025年10月1日

查看全文
  • Rui Li,
  • Debin Fang,
  • Pengyu Wang

摘要

Electric car-sharing systems (ECS) are increasingly adopting real-time pricing (RTP) to better respond to fluctuating demand and diversify service offerings for enhanced competitiveness. This study evaluates the performance of self-service and door-to-door ECS service strategies under RTP. To capture the complexities of real-world operations, we first construct realistic ECS networks by integrating point-of-interest data and graph neural network training. A duopoly game model is then developed, where two types of firms adopt distinct service strategies and adjust prices in real-time to maximize profits. Using deep reinforcement learning, the firms’ decisions are simulated and optimized across three ECS networks—Chongqing, Wuhan, and a synthetic network. Performance is evaluated across four key metrics: pricing, consumer utility, firm profit, and market share. The results reveal that door-to-door services outperform self-service models in terms of market share and profitability, as consumers prioritize convenience over price. The findings provide strategic insights for ECS operators.