Suggestions for Further Work

1. There is still some scope for further investigation into the optimisation procedure. As well as the three different schemes mentioned in section for minimising the variance, it has also been suggested [11] that a lower energy could be obtained by minimising a combination of the variance and the energy. This is currently being examined with the germanium atom to see if it produces significantly better results.
2. There is still the possibility of optimising the Slater determinant part of the trial wavefunction, . The overall scaling of these determinants is governed by the function, but there still remains the possibility of optimising the individual one-electron orbitals.
3. Optimising the Slater determinants offers the possibility of adjusting the nodes in the trial wavefunction, . As the fixed node approximation is one of the most likely candidates for the residual variance experienced during the optimisation of both the Jastrow and Chi terms in , releasing the nodes may also enable the variance to be significantly reduced.