next up previous contents
Next: A New u function Up: Optimising the u Function Previous: Results of Optimising the

Removing the tex2html_wrap_inline7307 Prefactor in tex2html_wrap_inline7145

One possible explanation for the new tex2html_wrap_inline7145 function of Eq.(gif) in the Jastrow factor being unable to reduce the energy and variance of energy significantly is that there are too many high powers in the prefactor of the Chebyshev expansion and these reduce its ability to produce good results. As the electron density close to the ion cores is very low compared to that in the bond (see Figure gif), it is possible that removing the tex2html_wrap_inline7307 prefactor to the Chebyshev expansion will not cause any instability. The argument here being that even though without this prefactor there is a cusp in tex2html_wrap_inline7145 when one electron moves through the ion, as the density close to the ion is so low, this is a very unlikely event.

The optimisations performed above were repeated for the tex2html_wrap_inline7145 function with no tex2html_wrap_inline7233 term in the prefactor.


The above assumption proved to be correct in that this did not appear to produce any serious instability into the optimisation. However, it also failed to produce any significant further reduction in either the energy or the variance of the energy.

Andrew Williamson
Tue Nov 19 17:11:34 GMT 1996