Maintenance Centered Service Parts Inventory Control Defended on Thursday, 30 May 2013
High-tech capital goods enable the production of many services and articles that have become a part of our daily lives. Examples include the refineries that produce the gasoline we put in our cars, the photolithography systems that enable the production of the chips in our cell phones and laptops, the trains and railway infrastructure that facilitate public transport and the aircraft that permit us to travel long distances. To prevent costly production disruptions of such systems when failures occur, it is crucial that service parts are readily available to replace any failed parts. However, service parts represent significant investments and failures are unpredictable, so it is unclear which parts should be stocked and in what quantity.
In this thesis, analytical models and solution methods are developed to aid companies in making this decision. Amongst other things, we analyze systems in which multiple parts need replacement after a failure, a situation that is frequently encountered in practice. This affects the ability to complete repairs in a timely fashion. We develop new modeling techniques in order to successfully apply scalable deterministic approaches, such as column generation techniques and sample average approximation methods, to this stochastic problem. This leads to solution techniques that, unlike traditional methods, can ensure that all parts needed to complete maintenance are readily available. The approach is capable of meeting the challenging requirements of a real-life repair shop.
Keywords
service parts; spare parts; inventory control; obsolescence; redundancy; intermittent demand forecasting; ATO systems; column generation; stochastic programming; stochastic dynamic programming