Business Process Simulation in Academia

Business Process Simulation

One of the main challenges is to create simulation models that accurately reflect the real-world process of interest. Moreover, we do not want to use simulation only to address strategic questions, but also for tactical and even operational decision making. Simulations can be used to predict performance under various circumstances, e.g., different business process reengineering alternatives can be compared with the current situation. The value of such predictions stands or falls with the quality of the simulation model.

Business process simulation involves developing an accurate simulation model which reflects the behavior of a process, including the data and resource perspectives, and then performing simulation experiments to better understand the effects of running that process. There are several steps involved in simulating a business processes. First, the business process is mapped onto a process model, possibly supplemented with process documentation facilities. Then, the sub processes and activities are identified. The control flow definition is created by identifying the entities that flow through the system and describing the connectors that link the different parts of the process. Lastly, the resources are identified and assigned to those activities where they are needed. The process model should be verified to ensure that the model does not contain errors. Before a business process can be simulated, the performance characteristics, such as throughput time and resource utilization, need to be considered. For statistically valid simulation results, a simulation run should consist of multiple sub runs, and each of these sub runs should have a sufficient run time. During the simulation, the simulation clock advances. The simulation tool may show an animated picture of the process flow or real-time fluctuations in the key performance measures. When the simulation is finished, the simulation results can be analyzed. To draw useful and correct conclusions from these results, statistical input and output data analysis is performed (Wynn, Dumas, & Fidge, 2007).