Design of a unit load AS / RS for work - in - process

Abstract

This thesis considers three problems concerned with designing a unit load automated storage/retrieval system (AS/RS) for work-in-process. The AS/RS can be characterized to be composed of a single rack and operated under a flexible manufacturing system (FMS) environment. The first problem is concerned with the optimal design of single-rack AS/RS with rectangular-in-time rack where a single storage/retrieval (S/R) machine is operated in dual command cycles. In the system, the randomized storage/retrieval policy is applied. The mid-point on the rack bottom line is then found as the optimal location of input/output (I/O) point for S/R machine which minimizes the expected dual command travel time. This is directed to the optimal rack shape search which is made under a constraint placed on the total number of rack openings to confirm that the optimal shape factor (relative time-measure of the rack height to bottom length) is about 0.60 in normalized rack. Finally the expected measure of the so-called no-cost zone is derived for a lower bound of the expected dual command travel time. The second problem is to determine the system size of the AS/RS. This problem is considered here for three different cases in the type of arrival items at storage space of AS/RS such as single type of items, multiple types of items, and single type of items with batch arrival. For each of these cases, the system size is measured as the number of storage spaces (inducing a storage capacity constraint) and the problem is analyzed by investigating its equivalent finite capacity queueing model or its equivalent closed queueing network model for which the long-run probability of the number of occupied storage spaces and the system blocking probability are derived. These probabilities are then used to construct a computational algorithm to search for the optimal system size by introducing a management-setting protection level representing the maximum blocking probability by which the management...