ISSN:
1573-2878
Keywords:
Two-point boundary-value problems
;
difference equations
;
invariant imbedding
;
Riccati transformation
;
iterative methods
Source:
Springer Online Journal Archives 1860-2000
Topics:
Mathematics
Notes:
Abstract The paper proposes an iterative solution method for discrete-time, nonlinear, two-point boundary-value problems (TPBVP) of the form: $$\begin{gathered} x(k) - x(k - 1) = f(k, x(k - 1), p(k)), \hfill \\ p(k) - p(k - 1) = g(k, x(k - 1), p(k)), \hfill \\ \end{gathered} $$ subject to $$h(x(0), p(0)) = 0,e(x(N), p(N)) = 0.$$ It is a counterpart of a method recently proposed by the authors for similar continuous-time TPBVPs with ordinary differential equations. The method, based on invariant imbedding and a generalized Riccati transformation, reduces the TPBVP to a pair of approximate initial-value problems with ordinary difference equations. Numerical tests are run on two examples originating in optimal control problems.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1007/BF00932889
Permalink