A Hardware Architecture of Model Predictive Control for a Real-Time Scheduling Algorithm

Model predictive control or model-based prediction control is one of the most powerful tools to robust plant control. The controller, particularly in vital and sensitive processes, requires a robust optimization method of cost function such as Quadratic Programming which is very time-consuming. Packet Scheduling for data transmission networks is a useful way to achieve better QoS and bandwidth optimization. In this paper we present an efficient hardware implementation to make MPC calculations (Especially Quadratic Programming) faster and make it suitable to be utilized as a controller in queue scheduling systems. Simulation results show that our architecture is faster and contains less area compared to previous cases.