1: \begin{abstract}

2:     In this paper the isogeometric % implementation of the

3:     Nyström method is presented. It is a technique for the numerical

4:     solution of boundary integral equations. It's outstanding feature

5:     is that it requires only pointwise function evaluations on the

6:     boundary of the computational domain just like isogeometric

7:     collocation methods. The points can be mapped from local to global

8:     parameter space using any geometrical mapping method. Therefor it

9:     is suitable for subdivision surfaces, T-splines or tensor product

10:     surfaces described by non-uniform rational b-splines. %

11:     The method is thoroughly integrated into the isogeometric

12:     framework.

13:     % i.e. boundary representations from computer aided geometric

14:     % design are directly taken for the definition of the geometry and

15:     % the approximation.

16:     The regularization of the arising singular integrals is performed

17:     by means of local correction with \BEZ-splines. The presented

18:     isogeometric Nyström method is applied to the Laplace and the

19:     \Lame-Navier equations. Numerical tests show higher order

20:     convergence in two and three dimensions. Additionally, the method

21:     is also applied to practical problems. It is concluded that the

22:     presented approach provides a viable alternative to currently used

23:     methods for solving boundary integral equations, but has some

24:     limitations.

25:   \end{abstract}

26: