An efficient parallel framework for process discovery using OpenMP

A process model is a graphical representation of the actual business process that is being executed. To build a process model from an event log, process discovery algorithms are used which are complex in nature and require prolonged execution as they involve extraction of the various ordering relations that exist between the events present in that event log. Given the exponential increase of data in event log, it is significant to have a robust and effective implementation of the computation intensive process discovery algorithms through parallel computing to generate a process model. Motivated by this theme the present work proposes a parallel computing approach to implement the Alpha algorithm for process discovery using the OpenMP application programming interface (API). An appropriate parallel programming framework to reduce the execution time by exploiting parallelism at the level of data, as well as task through a thorough analysis of the steps involved in the Alpha algorithm, has been developed. The effectiveness of the developed approach is presented on the basis of speedup factor through several experiments. The highest and the lowest speedups achieved were 13.24x and 4.71x respectively.