Cache Enhanced Split-Point-Based Alignment Calculation

Abstract

The execution of (business) processes often deviates from their behavioral specification (e.g., captured in a BPMN model). Conformance checking techniques evaluate whether event logs, i.e., data records capturing process behavior, and process models conform to each other. As such, conformance checking techniques provide insights into the correctness of the process execution. Alignments are conformance checking artifacts used to compute conformance metrics and, particularly, diagnostics. Several alignment algorithms exist, yet most existing methods solve an underlying search problem in which one typically calculates a heuristic to guide the search. Recently, a promising novel search approach was presented that reduces the overall number of heuristics required to solve the alignment problem. This paper extends this approach by proposing a caching strategy that improves the overall search speed and efficiency. We conducted a large set of experiments, confirming that the overall search efficiency increases significantly due to our contribution.

Authors 2

1. Tian Li Aachen Chair of Process and Data Science

   RWTH Aachen University

   Affiliation as printed

   RWTH Aachen University,Chair of Process and Data Science,Aachen,Germany

   Chair of Process and Data Science, RWTH Aachen University, Aachen, Germany

2. Sebastiaan J. van Zelst Aachen Fraunhofer Institute for Applied Information Technology FIT Germany

   RWTH Aachen University · Fraunhofer Institute for Applied Information Technology

   Affiliation as printed

   RWTH Aachen University,Fraunhofer Institute for Applied Information Technology FIT Germany,Sankt Augustin,Germany

   Fraunhofer Institute for Applied Information Technology FIT Germany, RWTH Aachen University, Sankt Augustin, Germany