STRUCTURE OPTIMIZATION FOR MULTIPLE MODEL CONTROL

摘要

This paper deals with structure optimization of thernnetwork of linear local models. The proposed techniquernassumes initial equidistant partition of the operatingrnspace. The structure of the model is then optimisedrnthrough iterative algorithm that includes merging similarrnlocal models. The merging condition is based on therncomputation of Mean Squared Error and similarity of thernlocal models in closed loop given by the gap metric. Thernapproach is illustrated by a simulation study of a pHrnneutralization process.