þÿ<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN"> <HTML><head><meta name=Titelcontent="JBPC Vol. 6, 1, 2006 ABSTRACT"> <meta http-equiv="Content-Type" content="text/html; charset=utf-8"> <title>JBPC Vol. 6, 1, 2006 ABSTRACT </title></head> <body link="#0000FF"><center><h1><font color="#006600">The Journal of Biological Physics and Chemistry</font></h1></center> <p></p><p></p> <b><center>2006<p>Volume 6, Number 1, p.p. 83-89</center></b> <br> <div> <p><b><font size=+2> Reducing the computational cost of <i>ab initio</i> potential energy calculations by using a dual-level approach </font></b></p> <p> <b> Akio Kawano<sup>1</sup> and Gia G. Maisuradze<sup>2</sup> </b> <br><br> <i> <sup>1</sup> Holistic Simulation Research Program, the Earth Simulator Center, Japan Agency for Marine-Earth Science and Technology (JAMSTEC), 3173-25 Showa-machi, Kanazawa-ku, Yokohama, 236-0001, Japan <br> <sup>2</sup> Department of Chemistry and Chemical Physics Program, University of Nevada, Reno, NV 89557, USA</i> </p> <p> We present a dual-level approach (a simplified version of the dual-level method proposed by Nguyen et al.) combined with the interpolated moving least-squares (IMLS) and modified Shepard (MS) methods as a tool for the reduction of CPU time of <i>ab initio</i> calculations. It is shown that the dual-level approach combined with IMLS has a noticeable advantage over the single-level method. Hence the dual-level/IMLS combination allows us to obtain the fit with the same accuracy and save CPU time while using the data points calculated straightforwardly from the single-level <i>ab initio</i> calculations. We show that the dual-level/IMLS combination is more efficient than the dual-level/MS combination. We also propose a further approach combined with a modified scaled Gaussian-3 (G3SX) theory, which belongs to the category of dual-level methods, to reduce computational time and cost for <i>ab initio</i> calculations. We discuss the range of applicability of the methods presented. </p> <b>Keywords: </b> IMLS, modified Shepard, scaled Gaussian-3 theory, single-level approach <br></div><p></p> <center><p><i><font size=-1><a href="jbpc20606.html">back to contents</a></font></i></p></center> </body></html>