A neural network graph partitioning procedure for grid-based domain decomposition
✍ Scribed by C. C. Pain; C. R. E De Oliveira; A. J. H. Goddard
- Publisher
- John Wiley and Sons
- Year
- 1999
- Tongue
- English
- Weight
- 300 KB
- Volume
- 44
- Category
- Article
- ISSN
- 0029-5981
No coin nor oath required. For personal study only.
✦ Synopsis
This paper describes a neural network graph partitioning algorithm which partitions unstructured ÿnite element/volume meshes as a precursor to a parallel domain decomposition solution method. The algorithm works by ÿrst constructing a coarse graph approximation using an automatic graph coarsening method. The coarse graph is partitioned and the results are interpolated onto the original graph to initialize an optimization of the graph partition problem. In practice, a hierarchy of (usually more than two) graphs are used to help obtain the ÿnal graph partition. A mean ÿeld theorem neural network is used to perform all partition optimization. The partitioning method is applied to graphs derived from unstructured ÿnite element meshes and in this context it can be viewed as a multi-grid partitioning method.