Discovering optimized process model using rule discovery hybrid particle swarm optimization

Abstract

This paper presents a bio-inspired hybrid method which concentrate on the optimal or a near-optimal business process model from an event log. The discovery of Hybrid Particle Swarm Optimization (Hybrid PSO) algorithm comes from the combination of Particle Swarm Optimization (PSO) algorithm and Simulated Annealing (SA) method. This paper presents a method which combines Rule discovery task and Hybrid PSO. The proposed method can discover not only classification rules that produce the most optimal business process model from event logs, but also can optimize the quality of process model. To be formulated into an optimization problem, we use rule discovery task to get the high accuracy, comprehensibility and generalization performance. After we get the results from rule discovery task, we use Hybrid PSO to resolve the problem. In this proposed method, we use continuous data as data set and fitness function as evaluation criteria of quality of discovered business process model. As final results, we prove that the proposed method has the best results in terms of average fitness and number of iterations, compared with classical PSO algorithm and original hybrid PSO algorithm.