Mine Your Simulation Model: Automated Discovery of Business Process Simulation Models from Execution Data

Abstract

Business process simulation is a versatile technique to estimate the performance of a process under multiple scenarios. This capability allows analysts to compare alternative options to improve a business process. A common roadblock for business process simulation is the fact that constructing high-fidelity simulation models is cumbersome and error-prone.Modern information systems such as Enterprise Resource Planning or Customer Relationship Management systems store detailed execution logs of the business processes they support. These execution logs can be used to automatically discover simulation models. However, discovering high-accuracy simulation models from business process execution data turns out to be a challenging problem due to the numerous factors that affect the performance of real processes. One of the major challenges is accounting for various work patterns, including multitasking, task prioritization, batching, resource availability schedules, and time-varying resource performance (e.g. fatigue effects).In this talk, I will give an overview of recent research in the field of automated discovery of business process simulation models. I will outline two approaches: one that uses process mining, curve fitting, and Bayesian optimization to discover and enhance a process model from an event log, and another approach that combines process mining with deep learning techniques. I will discuss the relative merits of these approaches and sketch open research challenges and questions.