ABSTRACTManagement goals in distribution inventories are often expressed in terms of the maximum percentage of aggregate sales that should be back ordered. This paper compares several strategies for allocating total inventory investment to each item stocked in order to meet such goals. Computational results are given from a wholesale distribution inventory. The results show that multi‐item strategies (which consider the interactions between items) require substantially less investment to meet management goals than strategies that treat each line item independently. All models in this research are approximations based on the assumptions commonly used in practic