Multiplexed Model Predictive Control

Most academic control schemes for MIMO systems assume all the control variables are updated simultaneously. MPC outperformsother control strategies through its ability to deal withconstraints. This requires on-line optimization, hencecomputational complexity can become an issue when applying MPC tocomplex systems with fast response times. The multiplexed MPCscheme described in this paper solves the MPC problem for eachsubsystem sequentially, and updates subsystem controls as soon asthe solution is available, thus distributing the control movesover a complete update cycle. The resulting computational speed-upallows faster response to disturbances, and hence improvedperformance, despite finding sub-optimal solutions to the originalproblem. This paper presents initial stabilityresults for two variants of multiplexed MPC.