My Math Forum QUick converging mimimization without derivatives?
 User Name Remember Me? Password

 Calculus Calculus Math Forum

 March 27th, 2018, 11:40 AM #1 Newbie   Joined: Oct 2013 Posts: 20 Thanks: 0 QUick converging mimimization without derivatives? We have three points: $\displaystyle x_0, x+h$ and $\displaystyle x-h$. We interpolate parabola using Stirling formula. Next, x:=parabola extreme. We can much decrease h and have next 3 points:$\displaystyle x_0, x+h$ and $\displaystyle x-h$. One problem: fast decreasing accuracy because of zero in denominator. Code:  static double f(double x) { return (x - 1) * (x - 2) * (x - 3); } static void minparab(double x0,double h) { double kappa = 13.6; double xe = 1.4226497308103742355;//exact value for f(x) for (int i = 0; i < 10; i++) { double y0 = f(x0); double yp = f(x0 + h); double ym = f(x0 - h); double x = x0 + h * (0.25 * yp- 0.25 * ym) / (y0 - 0.5 *ym - 0.5 *yp); double hmin = Math.Abs(x - x0); if (hmin > h) h = hmin; double h2 = h * h / kappa; Console.WriteLine("{0}", x-xe); x0 = x; h = h2; } } static void test() { minparab(1.5, 0.1); }

 Thread Tools Display Modes Linear Mode

 Similar Threads Thread Thread Starter Forum Replies Last Post John Travolski Calculus 2 October 6th, 2016 09:26 AM ineedhelpnow Calculus 25 July 26th, 2014 10:02 PM bobcantor1983 Calculus 1 October 2nd, 2013 01:48 PM mathemadness Real Analysis 2 April 21st, 2010 10:41 AM milin Complex Analysis 8 October 19th, 2007 04:39 PM

 Contact - Home - Forums - Cryptocurrency Forum - Top

Copyright © 2019 My Math Forum. All rights reserved.