Event Data-Driven Feasibility Checking of Process Schedules

Numerous processes require dedicated scheduling of their to-be-executed activities. Various algorithms have been developed to computationally solve many different scheduling problems, allocating the available resources to predefined time slots of activity execution to (theoretically) maximize resource utilization efficiency. Yet, in industry, creating schedules for future process executions often remains a (primarily) manual, spreadsheet-based endeavor. Typically, manually created schedules are sub-optimal and potentially infeasible. At the same time, the event data stored in the information systems supporting the process can act as valuable input to further improve the general alignment of the schedule to the actual process execution. Therefore, in this paper, we propose a novel method that enables schedule feasibility checking based on historically recorded event data corresponding to the actual execution of the scheduled process. Our method serves as an input to detect significant issues in the project scheduling problems, which can be used to further improve the overall quality of the schedules computed. Our initial results confirm the general applicability of the proposed framework.