RLHGNN: Reinforcement Learning-driven Heterogeneous Graph Neural Network for Next Activity Prediction in Business Processes

2025-07-03

AI Summary

Predicting the next activity in business processes within service-oriented architectures requires jointly modeling sequential dependencies and non-sequential relationships (e.g., parallelism, conditions). Existing sequence models neglect structural constraints, while graph-based approaches suffer from homogeneous node representations and static topologies. To address this, we propose a heterogeneous graph neural network framework: (1) constructing a semantic heterogeneous process graph—derived from process mining—with three types of domain-specific edges; (2) integrating reinforcement learning (formulated as an MDP) to dynamically select optimal graph structures per process instance, enabling instance-adaptive modeling; and (3) designing a relation-specific heterogeneous graph convolution operator for aggregation. Evaluated on six real-world datasets, our method significantly outperforms state-of-the-art approaches, achieving ~1 ms inference latency—enabling real-time business monitoring and dynamic service composition.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Planning, Routing, and Scheduling: Activity and Plan Recognition
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Next activity prediction represents a fundamental challenge for optimizing business processes in service-oriented architectures such as microservices environments, distributed enterprise systems, and cloud-native platforms, which enables proactive resource allocation and dynamic service composition. Despite the prevalence of sequence-based methods, these approaches fail to capture non-sequential relationships that arise from parallel executions and conditional dependencies. Even though graph-based approaches address structural preservation, they suffer from homogeneous representations and static structures that apply uniform modeling strategies regardless of individual process complexity characteristics. To address these limitations, we introduce RLHGNN, a novel framework that transforms event logs into heterogeneous process graphs with three distinct edge types grounded in established process mining theory. Our approach creates four flexible graph structures by selectively combining these edges to accommodate different process complexities, and employs reinforcement learning formulated as a Markov Decision Process to automatically determine the optimal graph structure for each specific process instance. RLHGNN then applies heterogeneous graph convolution with relation-specific aggregation strategies to effectively predict the next activity. This adaptive methodology enables precise modeling of both sequential and non-sequential relationships in service interactions. Comprehensive evaluation on six real-world datasets demonstrates that RLHGNN consistently outperforms state-of-the-art approaches. Furthermore, it maintains an inference latency of approximately 1 ms per prediction, representing a highly practical solution suitable for real-time business process monitoring applications. The source code is available at https://github.com/Joker3993/RLHGNN.

Problem

• Predicting next activity in business processes for optimization
• Capturing non-sequential relationships in parallel executions
• Overcoming homogeneous representations in graph-based approaches

Innovation

• Transforms event logs into heterogeneous process graphs
• Uses reinforcement learning for optimal graph structure
• Applies relation-specific graph convolution strategies