Discovery and Simulation of Business Processes with Probabilistic Resource Availability Calendars

EasyChair Preprint 10778

Abstract

In the field of business process simulation, the availability of resources is captured by assigning a calendar to each resource, e.g., Monday-Friday 9:00-18:00. Resources are assumed to be always available to perform activities during their calendar. This assumption often does not hold due to interruptions, breaks, or because resources time-share across multiple processes. A simulation model that captures availability via crisp time slots (a resource is either on or off during a slot) does not capture these behaviors, leading to inaccuracies in the simulation output. This paper presents a simulation approach wherein resource availability is modeled probabilistically. In this approach, each availability time slot is associated with a probability, allowing us to capture, for example, that a resource is available on Fridays between 14:00-15:00 with 90% probability and between 17:00-18:00 with 50% probability. The paper proposes an algorithm to discover probabilistic availability calendars from event logs. An empirical evaluation shows that simulation models with probabilistic calendars discovered from event logs, replicate the temporal distribution of activity instances and cycle times of a process more closely than simulation models with crisp calendars.

Keyphrases: Business Process Simulation, Process Mining, resource modeling

@booklet{EasyChair:10778,
  author    = {Orlenys López-Pintado and Marlon Dumas},
  title     = {Discovery and Simulation of Business Processes with Probabilistic Resource Availability Calendars},
  howpublished = {EasyChair Preprint 10778},
  year      = {EasyChair, 2023}}