Abstract:As human-robot collaboration increases in the workforce, it becomes essential for human-robot teams to coordinate efficiently and intuitively. Traditional approaches for human-robot scheduling either utilize exact methods that are intractable for large-scale problems and struggle to account for stochastic, time varying human task performance, or application-specific heuristics that require expert domain knowledge to develop. We propose a deep learning-based framework, called HybridNet, combining a heterogeneous graph-based encoder with a recurrent schedule propagator for scheduling stochastic human-robot teams under upper- and lower-bound temporal constraints. The HybridNet's encoder leverages Heterogeneous Graph Attention Networks to model the initial environment and team dynamics while accounting for the constraints. By formulating task scheduling as a sequential decision-making process, the HybridNet's recurrent neural schedule propagator leverages Long Short-Term Memory (LSTM) models to propagate forward consequences of actions to carry out fast schedule generation, removing the need to interact with the environment between every task-agent pair selection. The resulting scheduling policy network provides a computationally lightweight yet highly expressive model that is end-to-end trainable via Reinforcement Learning algorithms. We develop a virtual task scheduling environment for mixed human-robot teams in a multi-round setting, capable of modeling the stochastic learning behaviors of human workers. Experimental results showed that HybridNet outperformed other human-robot scheduling solutions across problem sizes for both deterministic and stochastic human performance, with faster runtime compared to pure-GNN-based schedulers.

A Graph Attention Network Approach to Partitioned Scheduling in Real-Time Systems

Dynamic Job-Shop Scheduling via Graph Attention Networks and Deep Reinforcement Learning

Heterogeneous graph attention networks for scalable multi-robot scheduling with temporospatial constraints

Learning Topological Representations with Bidirectional Graph Attention Network for Solving Job Shop Scheduling Problem

Learning to schedule job-shop problems: representation and policy learning using graph neural network and reinforcement learning

Learning to Dynamically Coordinate Multi-Robot Teams in Graph Attention Networks

Mayfly Taylor Optimization-Based Graph Attention Network for Task Scheduling in Edge Computing

Multi-Agent Reinforcement Learning for Job Shop Scheduling in Dynamic Environments

Graph Embedding Based Wireless Link Scheduling with Few Training Samples

Network Contention-Aware Cluster Scheduling with Reinforcement Learning

A Data-Driven Approach to Dataflow-Aware Online Scheduling for Graph Neural Network Inference

Delay-Oriented Distributed Scheduling Using Graph Neural Networks

Dynamic job-shop scheduling using graph reinforcement learning with auxiliary strategy

Graph Neural Networks for Job Shop Scheduling Problems: A Survey

Learning Coordination Policies over Heterogeneous Graphs for Human-Robot Teams via Recurrent Neural Schedule Propagation

Petri-net-based deep reinforcement learning for real-time scheduling of automated manufacturing systems

Flexible job shop scheduling via deep reinforcement learning with meta-path-based heterogeneous graph neural network

A Graph-Based Neural Approach to Linear Sum Assignment Problems

Adaptive Scheduling for Machine Learning Tasks over Networks

Fast Real-Time Scheduling for Ethernet-Based Train Control Networks.

TempoKGAT: A Novel Graph Attention Network Approach for Temporal Graph Analysis