Coordinating from Memory: Graph-Structured Experience Reuse for Multi-Agent Adaptation in Dynamic Manufacturing

2026-07-22

AI Summary

This study addresses the challenge of coordination in multi-agent systems operating in dynamic manufacturing environments, where frequent disruptions—such as machine failures and urgent tasks—hinder effective collaboration, and existing approaches often neglect reusable historical coordination experiences. To overcome this limitation, the authors propose the Graph-Structured Experience Memory (GSEM) framework, which, for the first time, integrates heterogeneous relational graphs with graph neural networks to enable structured encoding and similarity-based retrieval of past experiences. This facilitates rapid policy adaptation across diverse disturbance types. Evaluated on dynamic flexible job shop scheduling tasks, GSEM consistently outperforms the strongest baseline, reducing makespan by 4.1%–10.0% and adaptation time by 33%–38%, with performance gains becoming more pronounced under higher disruption frequencies.

Abstract

Dynamic manufacturing environments require multi-agent systems to coordinate effectively under frequent operational disturbances such as machine failures, urgent job arrivals, and processing time variations. Existing multi-agent reinforcement learning approaches treat each disturbance episode independently, discarding valuable coordination experience that could accelerate future adaptation. In this paper, we propose a Graph-Structured Experiential Memory (GSEM) framework for multi-agent coordination in dynamic manufacturing. The framework encodes historical coordination episodes as heterogeneous relational graphs that capture task dependencies, machine states, and inter-agent collaboration patterns. When a new disturbance occurs, a graph neural network-based retrieval mechanism identifies structurally similar past episodes, enabling experience-guided policy adaptation rather than learning from scratch. Experiments on dynamic flexible job-shop scheduling benchmarks with three disturbance types show that GSEM reduces makespan by 4.1%-10.0% and adaptation time by 33%-38% compared to the strongest memory-augmented baseline, with the advantage increasing under higher disturbance frequency. Ablation studies and cross-disturbance transfer experiments further validate the necessity of graph-structured encoding and similarity-based retrieval and demonstrate the cross-disturbance generalizability of learned coordination patterns.

Problem

• multi-agent coordination
• dynamic manufacturing
• operational disturbances
• experience reuse
• adaptation

Innovation

• Graph-Structured Memory
• Multi-Agent Reinforcement Learning
• Experience Reuse
• Dynamic Manufacturing
• Graph Neural Networks