A Petri Nets based Generic Genetic Algorithm framework for resource optimization in business processes

Business process simulation (BPS) enables detailed analysis of resource allocation schemes prior to actually deploying and executing the processes. Although BPS has been widely researched in recent years, less attention has been devoted to intelligent optimization of resource allocation in business processes by exploiting simulation outputs. This paper endeavors to combine the power of a genetic algorithm (GA) in finding optimum resource allocation scheme and the benefits of the process simulation. Although GA has been successfully used for finding optimal resource allocation schemes in manufacturing processes, in this previous work the design of these algorithms is ad hoc, meaning that the chromosomes, crossover and selection operators, and fitness functions need to be manually tailored for each problem. In this research, we pioneer to design and implement a Petri Nets based Generic Genetic Algorithm (GGA) framework that can be used to optimize any given business processes which are modeled in Color Petri Nets (CPN). Specifically, the proposed GGA framework is capable of producing an optimized resource allocation scheme for any CPN process model, its task execution times, and the constraints on available resources. The effectiveness of the proposed framework was evaluated on archive management workflow at Macau Historical Archives and an insurance claim workflow from an Australian insurance company. In both case studies, the framework identified significantly improved resource allocation scheme relative to the one that existed when the data for the case studies were collected.