论文信息 - Applications of neural networks to fitting interatomic potential functions

Applications of neural networks to fitting interatomic potential functions

It is shown that neural networks can be used to fit a two-element many-body potential function. The system chosen is the C-H combination for which a many-body potential formulation due to Brenner exists. Comparison between this potential and the neural network indicates good agreement with both structure and energetics of the small C-H clusters and bulk carbon. However, because of the networks complicated structure, molecular dynamics simulations run at about a factor of 60-80% slower than with the Brenner many-body formalism.

[1] Steven Hobday,et al. Optimisation of carbon cluster geometry using a genetic algorithm , 1997 .

[2] Steven D. Brown,et al. Neural network models of potential energy surfaces , 1995 .

[3] J. Tersoff,et al. Empirical interatomic potential for carbon, with application to amorphous carbon. , 1988, Physical review letters.

[4] Roger Smith,et al. Early stages of bump formation on the surface of ion-bombarded graphite , 1995 .

[5] D. Brenner,et al. Empirical potential for hydrocarbons for use in simulating the chemical vapor deposition of diamond films. , 1990, Physical review. B, Condensed matter.

[6] Anders E. Carlsson,et al. Beyond Pair Potentials in Elemental Transition Metals and Semiconductors , 1990 .

[7] Brenner. Erratum: Empirical potential for hydrocarbons for use in simulating the chemical vapor deposition of diamond films , 1992, Physical review. B, Condensed matter.