|Institution:||University of Guelph|
|Keywords:||Mathematics ; Scheduling ; Hyper-heuristic ; Heuristic ; Job Shop Scheduling ; Evolutionary Computation ; Industrial Mathematics|
|Full text PDF:||https://atrium.lib.uoguelph.ca/xmlui/handle/10214/7825|
Since the automotive crisis of 2008 manufactures world wide have been forced to adopt leaner styles of day to day operation. This study focuses on modelling a real world scheduling problem for a high volume company which plays an integral part of the global supply chain. Two models are considered using rapidly changing demand from actual customers. In the first model a proof of concept is demonstrated on a subset of the over all data showing the benefits of specific low level heuristics. In the final model additional heuristics are employed to provide richer, robust solutions which easily meet or exceed the performance of highly trained employees that previously had been responsible for scheduling the facility.