Planning Production on an Unreliable Machine for Multiple Items Subject to Stochastic Demand

摘要: We develop an extension of the classical newsvendor model that incorporates multiple items, setup times, and unreliable machine. This is motivated by applications at metal stamping plants where machine reliability a key source uncertainty. Given fixed production schedule, finite horizon, known demand distribution, we formulate model, derive important properties this exploit these to provide solution algorithm determines cost minimizing quantities. Finally, present three simple extensions model: (1) method for rescheduling within planning (2) evaluate whether or not purchase option run overtime (3) permits modeling operates different speed depending on part being produced.