An approach for AI-based forecasting of maintenance orders for MRO scheduling

Abstract

Maintenance orders are difficult to plan and require a high degree of flexibility, because both the extent of the activity ("What needs to be done?"), the scheduling ("When is the repair to be carried out?") and spatial restrictions ("Where is the repair to be carried out?") are largely unknown at the beginning of an order. All this results in a wide-spread reactive maintenance coordination in the industry instead of an efficient proactive maintenance planning of the diverse process. This not only leads to losses in the form of waiting and downtimes, but the lack of transparency both in the utilization situation and about the status of each order leads to delivery date difficulties and wasted resources. In order to decisively improve the status quo, it is indispensable to improve the accuracy of information on the above-mentioned questions as soon as possible after receiving the order. In this paper, an approach for the application of AI in MRO scheduling is presented including the line of research which needs to be done to enable a holistic planning optimization with decision support.