arrow
Return

Sampling-Based Coordination-Informed Multiobjective Multirobot Reinforcement Learning

delete2026-08-13
delete0
PRE
AI
A
Antonio Marino
E
Esteban Restrepo
S
Soon‐Jo Chung
P
Paolo Robuffo Giordano
C
Claudio Pacchierotti
DOI:10.1109/tro.2026.3723950delete
deleteOriginal
deleteShare
deleteSave
Abstract

Abstract

En 中文
Multirobot systems must simultaneously optimize competing objectives while maintaining coordinated behavior. Existing multiagent reinforcement learning approaches often rely on fixed or centralized coordination, which limits adaptability and violates distributed constraints. This work introduces the coordination-informed multiobjective reinforcement learning (CIMORL) framework, integrating a distributed weight prediction mechanism, a privileged expert training strategy, and theoretical guarantees for Pareto-optimal solutions. We present the base CIMORL method alongside two sampling-based variants, CIMORL-tree search (TS) and CIMORL-model predictive path integral (MPPI), which leverage privileged global information during training to enable fully decentralized deployment. Experimental validation in cooperative and adversarial scenarios demonstrates a 21.2% hypervolume improvement and superior policy stability compared to state-of-the-art baselines. Real-world experiments with Crazyflie drones further validate the framework’s robustness in resource allocation and multiattacker multidefend scenarios under partial observability.
Keywords:
Distributed control
graph neural network
multiobjective optimization
reinforcement learning

Journal

IEEE Transactions on Robotics cover
IEEE Transactions on Robotics
IF:
10.5
Papers:
3.4K
Citations:
3.5W

Organization

C
California Institute of Technology
Scholars:
5.0K
Papers: 1.4K
Citations: 2.8W
U
University of Cambridge
Scholars:
6.8K
Papers: 2.6K
Citations: 7.1W
Cited Papers

Cited Papers

Generalized policy improvement for efficient and robust multi-objective reinforcement learning
err2026-03-04
err0
PREAI
errAlegre, Lucas N.; Bazzan, Ana L. C.; Roijers, Diederik M.; Nowe, Ann; da Silva, Bruno C.
errShare
errSave
errShare
errSave
errShare
errSave
Multi-agent deep reinforcement learning: a survey
err2021-04-15
err317
errOAAI
errGronauer, Sven; Diepold, Klaus
errShare
errSave
Monte Carlo tree search algorithms for risk-aware and multi-objective reinforcement learning
err2023-12-01
err0
PREAI
errHayes,Conor F.; Reymond,Mathieu; Roijers,Diederik M.; Howley,Enda; Mannion,Patrick
errShare
errSave
A Survey of Multi-Objective Sequential Decision-Making
err2013-10-18
err0
errOAAI
errD. M. Roijers; P. Vamplew; S. Whiteson; R. Dazeley
errShare
errSave
researcher View more