A relatively new approach which simplifies the management and control of inventory and also reduces inventory is component commonality. Due to its promising results both companies and academicians found interest in this approach. A common assumption made in previous studies is that the demands of the various products are independent. However, it is not uncommon for companies to offer products that are complementing or substituting each other. In this paper the assumption of independent demands is relaxed and the effect of products' demands correlation (positive or negative) on the performance and attractiveness of component commonality is studied. In order to obtain closed form solutions a correlation between uniform distributions is introduced. This work shows that when component commonality is being considered, demands' correlation should not be neglected due to its significant effect.

In this paper, we develop procedures for dividing rectangular warehouses into classes. We describe three types of classes and develop their corresponding traveling time models. A solution procedure involving a one-dimensional (or at most two-dimensional) search is developed. We show that the one-dimensional search procedure is very effective in solving most practical problems. Additionally, we show that most of the savings obtained by a full-turnover policy can be obtained with a relatively few classes-a result similar to the one obtained for the square AS/RS.

In order to improve their service level many automated storage/retrieval systems (AS/RS) have adopted a dual command policy. In a dual command policy both storage and retrieval of pallets are done within one roundtrip of the crane. The expected travel time for a dual command policy is composed of two one-way travel time elements used for storage and retrieval, and an interleaving time element used for moving from a storage location to a retrieval location. In this study we apply the nearest-neighbour policy to a class-based warehouse in order to decrease the interleaving time. We develop the methodology for determining the interleaving time for both square and rectangular (time-wise) warehouses. We show that using the nearest-neighbour policy in a class-based environment can decrease the interleaving time, resulting in a significant increase in the total throughput.