Robust stability of GPC as applied to a first order model with delay

The proof of robust stability of generalised predictive control (GPC) when applied to a process represented by a first-order model is considered in this paper. The method shows how under certain assumptions, the robustness margin in the presence of model-plant mismatch can be improved by appropriate adjustments of the GPC tuning parameters. These results are derived using a combination of Maple (1981-1990), a symbolic mathprocessor, and standard algebraic methods.