In this paper, we propose an evolutionary computing approach based on Genetic Algorithms for composing an efficient trace given a desirable utility function based on the observations made in the event logs of several peer-organizations. Our proposed approach works with a set of event logs from different peer-organizations and generates an efficient trace according to a utility function. The main advantage of our approach is that we primarily work with event logs that are more accurate representations of the actual execution of a process within an organization. Furthermore, we generate efficient traces that are put together through the identification of sub-parts of the observed traces that are locally optimal. We report on our experiments on the BPIC’15 dataset that show improvement in terms of the optimality of the generated traces.

Keywords