The joint replenishment problem: Optimal policy and exact evaluation method

• We characterize the optimal joint replenishment policy. propose a new, exact method to evaluate any Our vastly reduces state-space size when compared Markov approach. introduce generalized can-order policy and improve decomposition new stationary under compound Poisson demand. The makes use of an embedded chain that only considers state system after order is placed. resulting space reduction allows analysis instances until now could be evaluated using approximation procedures. In addition, not affected if we include nonzero lead times, backlog, lost sales. policy, these characteristics develop greedy-optimal algorithm generalizes well-known family in class policies. numerically show this marginally improves best conventional For sizeable systems with multiple items, latter can found our Markov-chain method. Finally, extend