Towards Learning-Based Distributed Task Allocation Approach for Multi-Robot System

This paper introduces a novel application of Graph Convolutional Networks (GCNs) for enhancing the efficiency of the Consensus-Based Bundle Algorithm (CBBA) in multi-robot task allocation scenarios. The proposed approach in this research lies in the integration of a learning-based strategy to approximate the heuristic methods traditionally used for scoring in the CBBA framework. By employing GCNs, the proposed methodology aims to learn and predict the score function, which is crucial for task allocation decisions in multi-robot systems. This approach not only streamlines the allocation process but also potentially improves the accuracy and efficiency of task distribution among robots. The paper presents a detailed exploration of how GCNs can be effectively tailored for this specific application, along with results demonstrating the advantages of this learning-based approach over conventional heuristic methods in various simulated multi-robot task allocation scenarios.