Adaptive finite-element method for electronic-structure calculations.

We present a method which combines the finite-element method with the adaptive curvilinear coordinates, and the method is applied to the electronic-structure calculation of a model potential system. Comparison with other real-space methods, such as the finite-difference method is also made, and the efficiency of our method is examined. In addition, it has some desirable properties, and will be useful as a part of the O(N) method for self-consistent calculations. \textcopyright{} 1996 The American Physical Society.