|Student Name:||Capt Theodore Shiveley|
|Thesis:||Determining Optimal Machine Replacement Events with Periodic Inspection Intervals|
|Location:||COA conference room|
|Date & Time:||03/06/2013 at 0930|
|Abstract:|| This research will examine the optimal maintenance and replacement policies for a generic machine with periodic inspection intervals. Markov Decision Processes (MDPs) will be used to govern both the inspection process and repair/replace decisions. The considered design will be a machine operating with an exponential lifetime distribution which can fail both spontaneously or found unserviceable at an inspection. Instead of the generally observed practice of allowing a machine to operate in one of several degradation states, we will instead focus on a machine which has only one operational state, but can fail in any user defined manner. The adopted policy will be one of choosing the actions with the projected lowest cost per unit time. The effects of infinite and finite time horizons will be taken into consideration to observe if a different policy is necessitated by that choice. Finally, empirical cumulative distribution functions of time until first replacement will also be constructed. The goal of this is to create a flexible framework which an operator may input the parameters of his machine into and in turn give them a forecast of what expected cost and maintenance schedules might be.