Abstract
Operating room scheduling is often done in steps. First, surgeries are assigned to an operating room's time blocks. Assigned surgeries are then sequenced. Idle time is often reserved at the end of the time block in order to buffer against possible overtime. This research focuses on the next step of determining the amount of time reserved for each of the pre-sequenced surgeries so that surgical teams know their exact start times. In this way the buffer time is redistributed to each of the surgeries in order to minimize total overtime and idling costs. The problem is modeled as a special periodic review inventory model and a simulation-based response surface method is used to optimize surgery start times for a single operating room with stochastic operation durations represented by an infinite set of stochastic scenarios. This purposed method does not require extensive computational effort and is easy for practitioners to implement.