Using Inductive Reasoning to Find the Cause of Process Delays

Abstract

Delays in business processes can have negative consequences for organizations, such as extra costs, missed deadlines, poor service, etc. Although they are easy to detect, it may be hard to find the actual reason for such delays. In this work we develop an approach to find the cause of delays based on the information recorded in an event log. The approach is based on a logic representation of the event log and on the application of decision tree induction to separate process instances according to their duration. In this decision tree, the delayed instances immediately stand out, and by following the path in the tree it is possible to extract a rule that characterizes those instances and therefore provides a possible explanation for the delay. We illustrate the approach in a set of experiments with event logs that are generated by simulation of a purchase process. In each experiment, there is a different cause of delay, and the approach succeeds in finding that cause. Additional experiments show that the approach is scalable and tolerant to noise.