Abstract
As the increasing complexity of semiconductor manufacturing, AMHS plays an important role to deal with the transport and storage of wafer lots in a 300mm wafer fab. As the tool layout and AMHS track design was confirmed, the next phase of AMHS design is to determine the optimum number of vehicle fleet size and stocker bin capacity with refer to transport demand and performance measure request. Finally, to deploy and implement the material control system (MCS) to manage AMHS vehicle management control (e.g. vehicle dispatching, vehicle allocation, vehicle types, vehicle routing, and congestion/deadlock control) for daily operations. In order to improve the transport performance, fab managers continuously modify MCS vehicle management method to reduce vehicle waiting time and transport time of the wafer lot. Many researchers have paid efforts to propose the approaches in vehicle dispatching, vehicle routing, and congestion or deadlock control. However, there are few researches to address the other MCS vehicle management problem, e.g. vehicle allocation control. In a large-scale wafer fab, there are many intrabays and over hundreds of vehicle moving within or between intrabays. With the uncertainty involved in wafer lot movement, dynamically allocating vehicles to each intrabay is very difficult. Furthermore, in the current vehicle dispatching system, MCS only dispatches one vehicle for one load port at a time but this method cannot reduce the vehicle waiting time for one load port which has more than two transport jobs to move by AMHS. In this dissertation, two vehicle management strategies was proposed to reduce the vehicle waiting time of the wafer lot and improve AMHS system effectiveness. In the first study, a dynamic vehicle allocation control (DVAC) method is proposed for a connecting transport AMHS in a 300mm wafer fab. The objective is to minimize the sum of the expected long-run average transport job waiting cost. An interesting exhaustive structure in the optimal vehicle allocation control is found in accordance with the Markov decision model. Based on this exhaustive structure, an efficient algorithm is developed to solve the vehicle allocation control problem numerically. The performance of the proposed method is verified by a simulation study. Compared with other methods, the DVAC method can significantly reduce the waiting cost of wafer lots for AMHS vehicle transportation. The control policy can help fab managers to enhance MCS vehicle control logic to reduce the empty vehicle arrival time. In the second study, a pre-dispatching vehicle (PDV) method is proposed to simultaneously call several empty vehicles to move to a load port for executing transport jobs at a time. In semiconductor manufacturing, a furnace tool has a long processing time of 6 h to 12 h, thus making its operation a form-batch manufacturing step. With this application, wafer lots are temporarily stored in stockers, and delivery to furnace tools for processing is executed when the work-in-process (WIP) level reaches four to six lots. Unlike current methods, where empty vehicles are sequentially assigned for movement to load ports, the PDV method simultaneously calls several empty vehicles to move to a load port with refer to available empty vehicles and pre-dispatching vehicle quantity. This finding is useful for fab managers to explore the possibility of applying the PDV method to other areas and to continuously reduce the vehicle waiting time of the wafer lot.