WH-EA: An Evolutionary Algorithm for Wiener-Hammerstein System Identification

Table 5

Performance measurements on benchmark data (SYSID’09). All the values are shown in mV. indicates the number of parameters used for the model.

Method/technique	(mV)
Nonparametric BLA, QBLA [41]	0.278	44
Classification of poles and zeros using QBLA [42]	0.286	26
Fractional model parameterization [43]	0.295	26
Advanced method [44, 45]	0.30	64
WH-EA (this paper)	0.306	26
Brute force method [45]	0.31	30
Scanning technique [46]	0.370	-
Polynomial nonlinear state space [47]	0.42	797
Generalized Hammerstein-Wiener [48]	0.481	47
Incremental nonlinear optimization [49]	0.679	25
LS-SVMs [50]	4.070	-
Biosocial culture [51]	8.546	34