Article ID: | iaor2000443 |
Country: | Japan |
Volume: | 46 |
Issue: | 2 |
Start Page Number: | 359 |
End Page Number: | 381 |
Publication Date: | Dec 1998 |
Journal: | Proceedings of the Institute of Statistical Mathematics |
Authors: | Tanemura Masaharu |
Keywords: | biology, geography & environment |
In this paper, problems of evenly spaced configuration of points on the sphere are discussed. After introducing the minimax optimal problem and the minimum energy problem on the sphere, a new method of obtaining an optimal configuration is proposed under an utterly different principle. The new method is called a spherical adjustment method. It is first shown that the spherical adjustment method is useful for generating a global optimal configuration of a small number of points. Then, it is shown that the new method is also useful for generating local optimal configurations of points for a wide range of the number of points. It is also shown that, by carefully selecting a good configuration as an initial condition, our new method is able to improve the equilibrium configurations of points, for instance, which are generated by Markov Chain Monte Carlo method under a repulsive interaction potential.