A potential energy function for the water dimer has been developed with an artificial neural network (back propagation of error algorithm). The potential energy surface was obtained with 6s3p3d/3s3p MP2 ab initio MO calculations. The trained neural network reproduced the potential energy surface of the water dimer very well, not only in the low-energy region but also in the high-energy region.
Bibliographical noteFunding Information:
This work was supported by research grants from the Basic Science Research Inst. Program of the Ministry of Education (BSRI-96-3448), Korea, from the US National Institutes of Health (GM-14312), and from the US National Science Foundation (MCB95-13167 and INT93-06345). We thank Professor K.S. Kim for providing the coefficients of the basis functions.
All Science Journal Classification (ASJC) codes
- Physics and Astronomy(all)
- Physical and Theoretical Chemistry