Spin glass, the travelling salesman problem, neural networks and all that

Venkataraman, G. ; Athithan, G. (1991) Spin glass, the travelling salesman problem, neural networks and all that Pramana - Journal of Physics, 36 (1). pp. 1-77. ISSN 0304-4289

[img]
Preview
PDF - Publisher Version
6MB

Official URL: http://www.ias.ac.in/j_archive/pramana/36/1/1-77/v...

Related URL: http://dx.doi.org/10.1007/BF02846491

Abstract

This paper presents an overview of diverse topics that are seemingly different but interrelated, with strong connections to statistical mechanics on the one hand and spin glass physics on the other. Written primarily for an inter-disciplinary audience, we start with a brief recapitulation of the relevant aspects of statistical mechanics, particularly those needed for understanding the recently-popular simulated-annealing technique used in optimization studies. Then follows a survey of the spin glass problem, with particular attention to the consequences of quenched randomness. The travelling-salesman problem is considered next, as also the impact made on it by the spin glass problem. Several examples are then presented of optimization studies wherein the simulated-annealing concept has been profitably used. Attention is also drawn in this context to the lessons provided by the spin glass problem. Finally, a brief survey of neural networks is made, essentially from a physicist's point of view. The different learning schemes proposed are discussed, and the relevance of spin models and their statistical mechanics is also discussed.

Item Type:Article
Source:Copyright of this article belongs to Indian Academy of Sciences.
Keywords:Spin Glass; Travelling Salesman Problem; Neural Networks; Simulated Annealing
ID Code:55762
Deposited On:19 Aug 2011 03:14
Last Modified:18 May 2016 07:53

Repository Staff Only: item control page