Multivariable Constrained Generalised Predictive Control; A Comparison with QDMC

A multivariable constrained Generalised Predictive Control algorithm is proposed. The solution of the multivariable constrained optimisation is achieved using an orthogonal matrix decomposition technique, viz. singular value decomposition. Apart from enhancing the numerical integrity of the solution, the decomposition also imparts useful information which could be used to automatically improve the behaviour of the closed loop system. The proposed algorithm was applied to a simulated liquid mixing system, and the results were observed to compare well with those obtained using Quadratic Dynamic Matrix Control.