Bayesian Inventory and Replacement Models Using Real-Time Condition Monitoring Information

点击翻译按钮获取中文摘要

Bayesian Inventory and Replacement Models Using Real-Time Condition Monitoring InformationAbstractThis grant provides funding to develop methods for incorporating real-time sensor information obtained through condition monitoring (CM) into inventory management and replacement decisions for service parts, with the goal of improving the management of service parts inventories. This research considers the problem faced by a manufacturer who manages inventory for a machine part that is used on a large number of geographically-dispersed machines and which is subject to deterioration. The degradation signal generated by the deterioration of an individual part, and captured via CM, is modeled using a Wiener process. This degradation model, which can be used to predict part failure, as well as demand for service parts, is periodically updated as new sensor information is obtained. A stochastic inventory model that incorporates the updated demand distribution will be developed and the form of the optimal inventory control policy will be studied. While adaptive inventory policies such as those developed in this research can help manufacturers to increase machine availability and reduce inventory costs, their implementation can be computationally intensive and may require the retention of a significant quantity of sensor data. Thus, this research will focus on the development of computationally tractable methods that take advantage of the available sensor information.If successful, this research will develop implementable and broadly applicable tools that will be of practical value to manufacturing firms, assisting them in taking advantage of CM technology to improve their after-sales service and to compete more effectively. This research also seeks to contribute to the development of essential linkages between condition monitoring technology, part maintenance and replacement, and inventory management. This research identifies a critical area for interdisciplinary collaboration, seeking to build connections between research on CM technology and research on the development of decision models that make use of information obtained through CM.