In this paper, the approximation problem of SISO fuzzy systems is discussed. Based on fact that can be represented by a linear combination basic functions (FBF's), we first give systematic and detailed analysis FBF's present five properties FBF's: structure similarity compatibility between membership FBF's, complementarity less fuzziness composition systems. These provide clear picture shape features FBF's. these obtain some systems: property, uniform convergent property universal property....

In this paper, the approximation properties of MIMO fuzzy systems generated by product inference are discussed. We first give an analysis basic functions (FBF's) and present several FBF's. Based on these FBF's, we obtain systems: 1) property which reveals mechanism systems; 2) uniform bounds between desired (control or decision) 3) convergent shows that with defined accuracy can always be obtained dividing input space into finer regions; 4) universal approximators extends some previous...

This paper establishes the approximation error bounds for various classes of fuzzy systems (i.e., generated by different inferential and defuzzification methods). Based on these bounds, accuracy is analyzed compared. It seen that class product inference center-average defuzzifier has better properties than min defuzzifier, defuzzified MoM defuzzifier. In addition, it proved can represent any linear multilinear function explicit expressions method are given.

Abstract In this paper the problem of traffic control during rush hour is tackled using optimal theory. A simple discrete-time model developed to describe dynamic behaviour oversaturated urban road networks. It seen that standard optimization techniques can not be used obtain trajectories for system due dimensionality difficulties which are particularly accentuated by hard inequality constraints on states and controls as well pure time delays arise in description. However, a recently...

A method is developed for providing the optimal feedback gain matrix high-order linear quadratic problems by a decentralized computational procedure. All calculations in this approach are done off-line. The resulting gains all initial conditions so that eventual on-line computation minimal. applicable to both regulator and servomechanism particularly attractive use infinite time case where even off-line small. 22nd-order numerical example used illustrate approach. system here practical one...

In this paper reduced-order modelling and control analysis of linear, discrete-time systems having dominant non-dominant modes are presented. Decoupling is achieved using an explicitly invertible linear transformation. A matrix norm condition derived, the satisfaction which enables approximate expressions for block-diagonalizing matrices, eigenvalue distribution state trajectories to be obtained. Design stabilizing feedback controllers developed it shown that two gain matrices needed...

Abstract In this paper, a preliminary study is made of the dynamic optimization problem for river with many polluters using recently developed model section Cam near Cambridge in England. Because high dimensionality system it computationally prohibitive to obtain optimal solutions. However, has fairly simple structure which could be utilized alleviate computational burden. A hierarchical proposed pollution control and strategy demonstrated on digital simulation. For practical implementation...

This paper presents the decomposition property of fuzzy systems using a simple, constructive, procedure. That is, by properly dividing input space into sub-input spaces, general system is decomposed several sub-fuzzy which are simplest in spaces. Based on systems, analysis can be divided two steps: first, analyze properties and then, use to extend results systems. Using this idea, applications given. The first application representation capability second class nonlinear control Then, based...

This paper presents a relationship between membership functions and approximation accuracy in fuzzy systems. suggests an idea to design such that the of systems is improved.

In this paper, a new approach is developed for the detection of malfunctioning sensors large-scale linear interconnected dynamical systems. The main idea used to do an overlapping decomposition in order design multiple decentralised observers. Differences between same state as estimated by different observers can be isolate sensors. A sixth-order ladder network example illustrate approach.

Abstract This paper deals with eigenvalue assignment in linear discrete control systems which have a two time-scale property. It has been shown that such can be decomposed into fast subsystem small eigenvalues and slow large eigenvalues. Separate is attained using independent feedback gains.

In this paper, a new approach is developed for the detection of malfunctioning sensors large-scale linear interconnected dynamical systems. The main idea used to do an overlapping decomposition in order design multiple decentralised observers. Differences between same state as estimated by different observers can be isolate sensors. A sixth-order ladder network example illustrate approach.