Optimization Control Of Single Conveyor-Serviced Production Station With Variable Service Rate Under Different Production Environment

Conveyor-Serviced Production Station (CSPS) is a typical automated manufacturing system, in which the parts are delivered to the limited capacity buffer by a conveyor and will be served in the production station. It is a very important abstract model in practical production, which appeared in the industrial production automation. The research of optimal control problem of CSPS system by finding out an optimal control policy to reduce the production wastage rate, obtain the optimal long-run expected cost of the system and improve production output is of great feasible significance.In the CSPS system, the production station can be regarded as a decision-making body named Agent and it can coordinate the parts’ arrival process and service process by the front control. The front control model is that the Agent determines whether there parts within the distance, which is in accordance with the current system state, if there at least one part within the distance, the Agent will wait for the first part to reach the station and place it into the buffer which is the arrival process, otherwise, the Agent will take a part from the buffer to serve, which is the service process. To reduce the loss of the parts and improve the performance of the system, the Agent need take a good balance between the parts’ arrival process and service process by controlling the distance in accordance with the state. The sojourn time of the pasts’ arrival process obey a random distribution follows the arrival rate, which can be used to indicate the speed of the arrival process and changed when the system environment changes, so the arrival rate can be used to describe the system environment; the sojourn time of the service process is subject to a random distribution which is controlled by the service rate, the service rate can be used to indicate the speed of the service process, i.e. the service rate reflects the size of the process intensity in the production process. In order to obtain an optimal control of the CSPS problems under different environmental issues, the arrival rate can be used to describe the different environments and the service rate should be controlled to adapt the different environments.This dissertation considers adding the service rate to the control decision variable under the cases that the arrival rate is fixed and changing, and a semi Markov decision process (SMDP) of the CSPS problem is given by regarding the arrive rate and the remaining amount of the buffer as the system state and regarding the look-ahead distance and service rate as the control variable, then the dynamic characteristics of SMDP model, policy iterative algorithm and Q(A) algorithm are given to solve the CSPS problem. Besides, the mutative arrival rate can be unknown to the Agent. In this case, the CSPS problem of the unknown arrival rate can also be modeled as a SMDP model, but for the unknown arrive rate, only learning algorithm can be given to solve this problem.Finally, the simulation results show that, in the case that the arrival rate is fixed, changeable and unknown, the control of the service rate has achieved good results and show the effectiveness of this method.