Abstract:
This thesis is concerned with design, analysis and simulation, of artificial
neural network based controllers for certain problems involving nonlinear control systems.
Consider the following two types of nonlinear systems namely (i) the internal
control systems and (ii) the boundary control systems.
1 Internal Control Systems
*(t) = f(x(t),u(t),t) x(0) = xo
where x(t) = d d t) , x(t) E R11 and u(t) E Rm are called state variable and control
variable respectively, f is a nonlinear function. Here the control function u appears
in the dynamical model itself.
2 Boundary Control Systems
x(t)=f(x) in Q = §1 x (0,T)
x(0)=x0 in S2 (2)
x=gx~o onE=Fx(0,T)
Here the control function g appears in the boundary condition. In this system S2 is a bounded domain (nonempty, open and connected) in R with a smooth boundary F.
Fo is nonempty and an open subset of the boundary F. g represents boundary control function, X; is the characteristic function of Eo = F0 x (0,T) .
