PM-EdgeMap brings real-time process mining to smart factory edges
Edge computing cuts latency for process mining in cyber-physical systems
Smart factories are evolving into Cyber-Physical Systems (CPS) that require real-time decision-making from sensor data. Process mining helps extract insights from event logs, but traditional cloud-based approaches introduce latency. In their paper 'The PM-EdgeMap: Towards Real-Time Process Mining on the Edge-Cloud Continuum,' Hendrik Reiter and colleagues at Kiel University address this by proposing a formal model for distributing process mining tasks across edge and cloud resources. They introduce PM-EdgeMap, which includes a formalism to describe datasets and the computing topology, enabling efficient allocation of mining operations (e.g., conformance checking) to nearby edge nodes.
The team validates their approach with a case study on an edge-based conformance checking algorithm. Results show that executing process mining algorithms at the edge significantly reduces response time while maintaining accuracy, making it viable for autonomous control loops in smart factories. This work bridges the gap between process mining and edge computing, offering a blueprint for real-time operational intelligence in manufacturing environments. The paper was presented at the BPM 2025 workshops and is available on arXiv (2606.12103).
- Formal model (PM-EdgeMap) defines how to distribute process mining datasets and algorithms across edge and cloud tiers.
- Case study uses edge-based conformance checking—proves latency reduction for real-time autonomous factory control.
- Framework targets cyber-physical systems, enabling near-instant insights from sensor data without cloud round-trips.
Why It Matters
Enables real-time process mining at the edge, slashing latency for autonomous decision-making in smart factories.