Skip to main navigation Skip to search Skip to main content

Multi-agent Soft Actor-Critic with Coordinated Loss for Autonomous Mobility-on-Demand Fleet Control

  • Technical University of Munich

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

We study a sequential decision-making problem for a profit-maximizing operator of an autonomous mobility-on-demand system. Optimizing a central operator’s vehicle-to-request dispatching policy requires efficient and effective fleet control strategies. To this end, we employ a multi-agent Soft Actor-Critic algorithm combined with weighted bipartite matching. We propose a novel vehicle-based algorithm architecture and adapt the critic’s loss function to appropriately consider coordinated actions. Furthermore, we extend our algorithm to incorporate rebalancing capabilities. Through numerical experiments, we show that our approach outperforms state-of-the-art benchmarks by up to 12.9% for dispatching and up to 38.9% with integrated rebalancing.

Original languageEnglish
Title of host publicationLearning and Intelligent Optimization - 19th International Conference, LION 19 2025, Proceedings
EditorsYingqian Zhang, Milan Hladik, Hossein Moosaei
PublisherSpringer Science and Business Media Deutschland GmbH
Pages1-15
Number of pages15
ISBN (Print)9783032091550
DOIs
StatePublished - 2026
Event19th International Conference on Learning and Intelligent Optimization, LION 2025 - Prague, Czech Republic
Duration: 15 Jun 202519 Jun 2025

Publication series

NameLecture Notes in Computer Science
Volume15744 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference19th International Conference on Learning and Intelligent Optimization, LION 2025
Country/TerritoryCzech Republic
CityPrague
Period15/06/2519/06/25

Keywords

  • autonomous mobility on demand
  • coordinated loss
  • deep reinforcement learning
  • hybrid learning and optimization
  • multi-agent learning

Fingerprint

Dive into the research topics of 'Multi-agent Soft Actor-Critic with Coordinated Loss for Autonomous Mobility-on-Demand Fleet Control'. Together they form a unique fingerprint.

Cite this