Abstract: This paper presents a Gaussian process model-based
short-term electric load forecasting. The Gaussian process model is
a nonparametric model and the output of the model has Gaussian
distribution with mean and variance. The multiple Gaussian process
models as every hour ahead predictors are used to forecast future
electric load demands up to 24 hours ahead in accordance with the
direct forecasting approach. The separable least-squares approach that
combines the linear least-squares method and genetic algorithm is
applied to train these Gaussian process models. Simulation results
are shown to demonstrate the effectiveness of the proposed electric
load forecasting.
Abstract: This paper presents a nonparametric identification of
continuous-time nonlinear systems by using a Gaussian process
(GP) model. The GP prior model is trained by artificial bee colony
algorithm. The nonlinear function of the objective system is estimated
as the predictive mean function of the GP, and the confidence
measure of the estimated nonlinear function is given by the predictive
covariance of the GP. The proposed identification method is applied
to modeling of a simplified electric power system. Simulation results
are shown to demonstrate the effectiveness of the proposed method.
Abstract: In this paper, we are concerned with the design and
its simulation studies of a modified extremum seeking control for
nonlinear systems. A standard extremum seeking control has a simple
structure, but it takes a long time to reach an optimal operating point.
We consider a modification of the standard extremum seeking control
which is aimed to reach the optimal operating point more speedily
than the standard one. In the modification, PD acceleration term
is added before an integrator making a principal control, so that it
enables the objects to be regulated to the optimal point smoothly. This
proposed method is applied to Monod and Williams-Otto models to
investigate its effectiveness. Numerical simulation results show that
this modified method can improve the time response to the optimal
operating point more speedily than the standard one.
Abstract: This paper presents a method of model selection and
identification of Hammerstein systems by hybridization of the genetic
algorithm (GA) and particle swarm optimization (PSO). An unknown
nonlinear static part to be estimated is approximately represented
by an automatic choosing function (ACF) model. The weighting
parameters of the ACF and the system parameters of the linear
dynamic part are estimated by the linear least-squares method. On
the other hand, the adjusting parameters of the ACF model structure
are properly selected by the hybrid algorithm of the GA and PSO,
where the Akaike information criterion is utilized as the evaluation
value function. Simulation results are shown to demonstrate the
effectiveness of the proposed hybrid algorithm.
Abstract: This paper deals with an on-line identification method
of continuous-time Hammerstein systems by using the radial basis
function (RBF) networks and immune algorithm (IA). An unknown
nonlinear static part to be estimated is approximately represented
by the RBF network. The IA is efficiently combined with the
recursive least-squares (RLS) method. The objective function for the
identification is regarded as the antigen. The candidates of the RBF
parameters such as the centers and widths are coded into binary bit
strings as the antibodies and searched by the IA. On the other hand,
the candidates of both the weighting parameters of the RBF network
and the system parameters of the linear dynamic part are updated
by the RLS method. Simulation results are shown to illustrate the
proposed method.