next up previous contents
Next: Wavelet-based method Up: Tree Method Previous: Tree Method

Panel clustering method

Panel clustering method proposed by Hackbusch and Nowak[38] gives an algorithm of the order of $O(N
\log^{d+2}N)$ with the help of an expansion of a kernel function and decompositions of the spatial domain with ``panels'' where d is a dimension of a problem. As one views the algorithm of panel clustering method one will see that panel clustering method is similar to FMM or the tree method. The use of panel clustering method has been investigated by several authors: by Hayami and Sauter[39,40,41] and Sauter[73].

Ken-ichi Yoshida