Linear approximation model network and its formation via evolutionary computation. (Q5955829)

To overcome the deficiency of local model network (LMN) techniques, the authors develop an alternative linear approximation model (LAM). Such a network can model a nonlinear system with multiple linear models fitted along operating trajectories. The linear models are valid for the entire operating trajectory. LAMs can evolve from sampled step response data directly, eliminating the need for local linearization. The structural difference between a LAM network and an LMN is that the overall model of the latter is a parameter-varying system and hence nonlinear in some sense while the former remains linear time-invariant (LTI). LAM is easy to use for control system design. Numerical examples are given.