Abstract:
We consider the Degree/Diameter problem for circulants – the problem of constructing large undirected circulant graphs (networks) with given degree and diameter. We develop a genetic algorithm for synthesis of large circulant graphs and implement its parallel version by supercomputer systems. The algorithm has found 28 new large circulant graphs which orders are better than the largest of the current known circulants from the record $(\Delta/D)$-circulant graphs table for degrees $12\le\Delta\le16$ and diameters $4\le D\le10$. Tab. 2, bibliogr. 29.