EPE Association: Pelincec 2005 - 07: Control Theory and Intelligent Control

Pelincec 2005 - 07: Control Theory and Intelligent Control
You are here: EPE Documents > 05 - EPE Supported Conference Proceedings > Pelincec 2005 > Pelincec 2005 - 07: Control Theory and Intelligent Control

   [return to parent folder]

   Optimal real-time control for dynamical systems under uncertainty
By R. Gabasov, F.M. Kirillova [View]
[Download]

Abstract: The synthesis problem for optimal control systems in the class of discrete control functions is under consideration. The problem is investigated by reducing to a linear programming (LP) problem with consequent use of dynamic version of the adaptive method of LP (Gabasov et al, 2000). Both perfect and imperfect information on behavior of the control system cases are studied. Methods and algorithms of operating the optimal controller, optimal estimators are described. For control system under uncertainty algorithms are obtained according to the principle of on-line control i.e optimal feedbacks to the systems are not constructed beforehand, their current values are supposed to be calculated in the course of control processes at every current moment and position or output. Results of computer experiments are given.

   Realization problem for positive multivariable discrete-time linear systems with delays in state vector and inputs
By Tadeusz Kaczorek [View]
[Download]

Abstract: The realization problem for positive multivariable discrete-time systems with delays in state and inputs is formulated and solved. Conditions for the solvability of the realization problem and the existence of a minimal positive realization are established. A procedure for computation of a positive realization of a proper rational matrix is presented and illustrated by examples.

   Robust stability of scalar positive discrete-time linear systems with delays
By Mikolaj Buslowicz [View]
[Download]

Abstract: Simple necessary and sufficient conditions for robust stability of scalar linear positive discrete-time systems with delays in the general case (non-linear uncertainty structure) and in two special cases: 1) interval positive system, 2) positive system with linear or multilinear uncertainty structure, are given. The proposed conditions are compared with the suitable conditions for non-positive discrete-time systems.

   Avoiding parameter growth of TSK fuzzy models
By Jacek Kabziñski [View]
[Download]

Abstract: We propose two relatively simple and effective procedures for creating neuro-fuzzy Takagi-Sugeno-Kang model and for tuning of TSK model parameters together with the rule-base structure optimisation. The main advantage of the first method is that the initial structure and parameters are set properly, so we need a few training iterations for the neural network representation of our model to converge. In the second approach the most important is rule reduction procedure –annihilation and fusion incorporated in a genetic optimisation algorithm. Numerical examples are provided.

Pelincec 2005 - 07: Control Theory and Intelligent Control
You are here: EPE Documents > 05 - EPE Supported Conference Proceedings > Pelincec 2005 > Pelincec 2005 - 07: Control Theory and Intelligent Control
	[return to parent folder]

	Optimal real-time control for dynamical systems under uncertainty By R. Gabasov, F.M. Kirillova	[View] [Download]
Abstract: The synthesis problem for optimal control systems in the class of discrete control functions is under consideration. The problem is investigated by reducing to a linear programming (LP) problem with consequent use of dynamic version of the adaptive method of LP (Gabasov et al, 2000). Both perfect and imperfect information on behavior of the control system cases are studied. Methods and algorithms of operating the optimal controller, optimal estimators are described. For control system under uncertainty algorithms are obtained according to the principle of on-line control i.e optimal feedbacks to the systems are not constructed beforehand, their current values are supposed to be calculated in the course of control processes at every current moment and position or output. Results of computer experiments are given.

	Realization problem for positive multivariable discrete-time linear systems with delays in state vector and inputs By Tadeusz Kaczorek	[View] [Download]
Abstract: The realization problem for positive multivariable discrete-time systems with delays in state and inputs is formulated and solved. Conditions for the solvability of the realization problem and the existence of a minimal positive realization are established. A procedure for computation of a positive realization of a proper rational matrix is presented and illustrated by examples.

	Robust stability of scalar positive discrete-time linear systems with delays By Mikolaj Buslowicz	[View] [Download]
Abstract: Simple necessary and sufficient conditions for robust stability of scalar linear positive discrete-time systems with delays in the general case (non-linear uncertainty structure) and in two special cases: 1) interval positive system, 2) positive system with linear or multilinear uncertainty structure, are given. The proposed conditions are compared with the suitable conditions for non-positive discrete-time systems.

	Avoiding parameter growth of TSK fuzzy models By Jacek Kabziñski	[View] [Download]
Abstract: We propose two relatively simple and effective procedures for creating neuro-fuzzy Takagi-Sugeno-Kang model and for tuning of TSK model parameters together with the rule-base structure optimisation. The main advantage of the first method is that the initial structure and parameters are set properly, so we need a few training iterations for the neural network representation of our model to converge. In the second approach the most important is rule reduction procedure –annihilation and fusion incorporated in a genetic optimisation algorithm. Numerical examples are provided.