This thesis deals with an assembly production line system which utilizes automated guided vehicles(AGVs) for its material handling. It is assumed that AGV delivers parts from automated storage and retrieval system (AS/RS) to each work station and a conveyor transports sub-assemblies from a work station to its succeeding station. In this system, determination of unit load size of parts carried by AGV becomes an important factor due to its effects on the in-process inventory level, the number of AGV required and the operating cost of the system. A non-linear mathematical model is presented which considers all the cost elements and the system constraints. Based on the properties of an optimal solution, an implicit enumeration algorithm is developed. Also, a heuristic solution procedure is proposed for the case of a large upper bound. The validity of each algorithm is illustrated through numerical examples.