Automatic Synchronisation Detection in Petri Net Performance Models Derived from Location Tracking Data

Abstract

The inference of performance models from low-level location tracking traces provides a means to gain high-level insight into customer and/or resource flow in complex systems. In this context our earlier work presented a methodology for automatically constructing Petri Net performance models from location tracking data. However, the capturing of synchronisation between service centres – the natural expression of which is one of the most fundamental advantages of Petri nets as a modelling formalism – was not explicitly supported. In this paper, we introduce mechanisms for automatically detecting and incorporating synchronisation into our existing methodology. We present a case study based on synthetic location tracking data where the derived synchronisation detection mechanism is applied

PEPERCORN: Inferring Performance Models from Location Tracking Data

“…This tool has been evaluated through a number of case studies in [1][2][3]. These case studies, conducted using synthetic location tracking data generated by LocTrackJINQS [6], employ several types of customer-processing systems, including systems with synchronisation, multiple customer classes and service cycles.…”

“…PEPERCORN is a Java-based implementation of our earlier work [1][2][3] which presented a methodology, based on a four-stage data processing pipeline (cf. Figure 1), that allows the automated construction of CGSPN performance models from high-precision location tracking data.…”