Wil M.P. van der Aalst
Department of Mathematics and Computer Science, Eindhoven University of Technology, P.O. Box 513, NL-5600 MB, The Netherlands email@example.com Abstract. Computer simulation attempts to “mimic” real-life or hypothetical behavior on a computer to see how processes or systems can be improved and to predict their performance under diﬀerent circumstances. Simulation has been successfully applied in many disciplines and is considered to be a relevant and highly applicable tool in Business Process Management (BPM). Unfortunately, in reality the use of simulation is limited. Few organizations actively use simulation. Even organizations that purchase simulation software (stand-alone or embedded in some BPM suite), typically fail to use it continuously over an extended period. This keynote paper highlights some of the problems causing the limited adoption of simulation. For example, simulation models tend to oversimplify the modeling of people working part-time on a process. Also simulation studies typically focus on the steady-state behavior of business processes while managers are more interested in short-term results (a “fast forward button” into the future) for operational decision making. This paper will point out innovative simulation approaches leveraging on recent breakthroughs in process mining.
Limitations of Traditional Simulation Approaches
Simulation was one of the ﬁrst applications of computers. The term “Monte Carlo simulation” was ﬁrst coined in the Manhattan Project during World War II, because of the similarity of statistical simulation to games of chance played in the Monte Carlo Casino. This illustrates that that already in the 1940s people were using computers to simulate processes (in this case to investigate the eﬀects of nuclear explosions). Later Monte Carlo methods were used in all kinds of other domains ranging from ﬁnance and telecommunications to games and workﬂow management. For example, note that the inﬂuential and well-known programming language Simula, developed in the 1960s, was designed for simulation. Simulation has become one of the standard analysis techniques used in the context of operation research and operations management. Simulation is particularly attractive since it is versatile, imposes few constraints, and produces results that are relatively easy to interpret. Analytical techniques have other advantages but typically impose additional constraints and are not as easy to use . Therefore, it is no surprise that in the context of Business Process Management (BPM), simulation is one of the most established analysis techniques supported by a vast array of tools.
people machines business processes
gather simulation results to answer “what-if” questions
gather data and model by hand
Fig. 1. Classical view on simulation: focus is on steady-state and model is made by hand.
Figure 1 positions business process simulation in the context of a “world” supported by information systems. In the “world” consisting of people, organizations, products, processes, machines, etc. information systems play an increasingly dominant role. Moreover, there is continuous need for process improvements resulting in a better performance (e.g., better response times, less costs, higher service levels, etc.). Simulation can assist in this. Figure 1 shows the traditional use of simulation were data is gathered and used to parameterize hand-made models. These models are then used for simulation experiments answering “what-if” questions. For simulating business processes at least three perspectives need to be modeled: (a) control-ﬂow, (b) data/rules, and (c) resource/organization. The control-ﬂow perspective is concerned with the ordering of activities and uses design artifacts such as...