Research on Online Anomaly Detection of Business Processes Based on Digital Twin-Intelligent Body Collaborative Architecture

The increasing complexity of business processes makes traditional rigid structure business process management methods difficult to apply in complex scenarios. This study establishes a collaborative integration framework of twin-intelligent bodies, combining digital twin and intelligent body technologies, based on dynamic declarative constraints. The digital twin enhances agents’ perception and decision-making, allowing real-time constraint adjustments. An association rule algorithm mines multi-perspective business process data to detect anomalies. Experiments on the DCGAN platform show an F1 score of 0.8–1 and a false alarm rate under 13%. The method outperforms LSTM and Transformer in comparative tests. The case study validated the effectiveness of the proposed model in real-world scenarios.