Tensor Product Model Transformation in Polytopic Model-Based Control

i

i

i

i

i

i

i

i

Chapter 9

Computational Relaxed TP

Model Transformation

Any method that operates on multidimensional data suffers from the curse of di-

mensionality, that is, complexity increases exponentially but the algorithms are not

scalable. Such a problem arises in the tensor product (TP) model transformation as

well when the model depends on many parameters. The size of the sampled model

increases exponentially with respect to the number of parameters, and the higher-

order singular value decomposition (HOSVD) cannot cope with the huge amount of

data.

A general solution to this problem is, of course, to enhance the available com-

putational resources.