An easy derivation of the order level optimality condition for inventory systems with backordering

We analyze the classical inventory model with backordering, where the inventory position is controlled by an order level, order quantity policy. The cost for a backorder contains a fixed and a time-proportional component. Demand can be driven by any discrete process. Order lead times may be stochastic and orders are allowed to cross. The optimality condition for the order-level, given some predetermined order quantity, is derived using an easy and insightful marginal cost analysis. It is further shown how this condition can easily be (approximately) rewritten in well-known forms for special cases.