Paper Type
Short
Paper Number
1600
Description
Shared bike rebalancing is critical for the operation and management of bike-sharing systems (BSSs) as a shared resource balancing problem in the sharing economy. This study proposes a two-step approach: optimizing the service level and rebalancing the shared bikes. To optimize the service level, we model bike pickups and returns as a random process, using queuing theory to determine the optimal service time window and the number of bikes to relocate. For bike rebalancing, we formulate it as a Markov decision process for large-scale stations, aiming to minimize travel times and the number of vehicles used. We design a multi-vehicle reinforcement learning method to find optimal solutions. The proposed algorithm serves as a general framework applicable to various BSSs and other shared resource balancing scenarios.
Recommended Citation
Feng, Jiahui; Li, Yingbo; and Liu, Hefu, "Service level optimizing and shared bike rebalancing based on multi-agent deep reinforcement learning" (2024). PACIS 2024 Proceedings. 20.
https://aisel.aisnet.org/pacis2024/track01_aibussoc/track01_aibussoc/20
Service level optimizing and shared bike rebalancing based on multi-agent deep reinforcement learning
Shared bike rebalancing is critical for the operation and management of bike-sharing systems (BSSs) as a shared resource balancing problem in the sharing economy. This study proposes a two-step approach: optimizing the service level and rebalancing the shared bikes. To optimize the service level, we model bike pickups and returns as a random process, using queuing theory to determine the optimal service time window and the number of bikes to relocate. For bike rebalancing, we formulate it as a Markov decision process for large-scale stations, aiming to minimize travel times and the number of vehicles used. We design a multi-vehicle reinforcement learning method to find optimal solutions. The proposed algorithm serves as a general framework applicable to various BSSs and other shared resource balancing scenarios.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.
Comments
AI