Workload Balancing and Cycle Time Reduction by Using Concurrent Performing of Dependent Tasks in BPMS

The emergence of business process management systems (BPMS) is a valuable achievement in information systems and is an outstanding meeting point of intelligent systems and information systems. BPMS provides the possibility to design, monitor, manage and improve the business processes. Organizations may adopt policies like cycle time reduction, production increase, production cost decrease, and so on. workload balancing may also be considered as an objective. Workload balancing guarantees fair work distribution among resources. In heavy load conditions or when the number of human resources is limited, workload balancing can increase the stability of the system. On the other hand, cycle time reduction would lead to waiting time reduction and customer satisfaction. In this paper, a novel approach is proposed to reduce cycle time of business processes by dispatching "dependent" tasks while maintaining workloads balanced. Dependent tasks are the tasks in a business process for which concurrent processing would help reduce cycle time. The concept of dependency is extracted by statistical analysis of event logs.To evaluate the proposed approach, BPI Challenge 2012 dataset is chosen to benchmark the three conducted experiments. The experiments shows that the proposed method (WB_FIFO+CPDT) outperforms WB_FIFO+DR and DSWB methods in cycle time reduction, especially in circumstances where arrival rate increases. The superiority of the proposed method in cycle time reduction is 16.6% in contrast to DSWB and 31.1% in contrast to WB_FIFO+DR.