Adaptive Scheduling Systems: A decision-theoretic approach

Abstract

This research mainly focuses on solving the dynamic job-shop problems where high uncertainties and problem complexity greatly affect the quality of schedules generated. This thesis proposes an adaptive scheduling system based on decision theory. Influence diagram is used as the main modeling tool in the approach. It facilitates the inclusion of all possible variables that influence the current shop-floor conditions into a single framework and also facilitates the inter-dependency among variables. The models are capable of addressing the effectiveness, reactivity, usability, and robustness of the schedules generated. Three approaches are proposed in the thesis, namely, proactive scheduling, reactive scheduling, and hybrid approach of proactive-reactive scheduling with periodic-event-driven review technique. The proposed system was tested and evaluated using a simulated shop-floor system. Methods to quantify the disruptions are also proposed. The results show that the system is capable of adaptively decides between proactive scheduling and reactive scheduling in the context of high shop floor uncertainties and possible disruptions.