Composite business processes: An evolutionary multi-objective optimization approach

Business process optimization has received little coverage compared to business process modeling and analysis techniques. This paper introduces composite business process models, i.e. conceptual business processes with tasks that each has its own library of alternatives. This paper formulates an optimization problem based on this concept. A series of experiments is designed to address processes of various sizes in terms of participating tasks and libraries of alternatives. Evolutionary algorithms such as NSGA2, SPEA2 and MOPSO attempt to generate optimum solutions to a bi-objective and tri-objective problem formulation. The results show that SPEA2 performs better in the bi-objective problem, while NSGA2 has a clear advantage in the tri-objective problem, although both provide good solutions in all instances. This paper attempts to establish a viewpoint regarding business processes that will provoke and encourage further optimization attempts in this area.