Generating valid reference Business Process model using genetic algorithm

With the growth of the Business Process Management (BPM) market, customer requirements based process customization has emerged as a key to competitive advantage. Where there are numerous demands for customization, a great number of business process model variations result. The existence of numerous process variations leads to both process redundancy and process underutilization, which impact on business performance in negative ways. Thus, there is a need to create a process reference model that can identify and find a representative model without redundancy. This paper introduces a new, Genetic Algorithm (GA)-based approach to generation of a valid business process reference model from a process repository. Previous research using the integer programming (IP) problem produced a reference process model. However, there are problems, pertaining to presentational limitations and unguaranteed validation processes. The GA procedure uses the randomness of genetic processes to resolve both problems in IP formulation. Near the end of this paper, we show the experimental results of the proposed method, which is conveniently executed using business process structure properties and the proximity of activities. It is believed that this process reference model can help a novice process designer to create a new process based on existing processes.