lec16_10182006 - 10.34, Numerical Methods Applied to...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon
10.34, Numerical Methods Applied to Chemical Engineering Professor William H. Green Lecture #16: Unconstrained Optimization. Unconstrained Optimization min x f(x ) x [k+1] x guess Require: x [k] f ( x [k+1] )<f(x [k] ) Which direction to move? Move Downhill Æ “Steepest Descent” very robust but poor convergence at the end x Figure 1. Diagram of steepest descent approach to global minimum. Unless you start on the center line, you will zigzag inefficiently going down contour lines is easy with this method f approx (x ) = f (x [k] ) + f | x [ k] ·(x –x [k] ) + ½(x –x [k] ) T B ·(x – x [k] ) (x – x [k] ) = p = - f (|| f || 2 / f T B f ) {Cauchy} B must be positive definite and not singular. p cauchy (- f /||f ||) Δ Å max step size allowed p steepest descent = p cauchy t a k e m i n | | p || Look at Figures 5.5 and 5.6 in BEERS Cite as: William Green, Jr., course materials for 10.34 Numerical Methods Applied to Chemical Engineering, Fall 2006. MIT OpenCourseWare (http://ocw.mit.edu), Massachusetts Institute of Technology. Downloaded on [DD Month YYYY].
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
Figure 2. An example of poor scaling. 10.34, Numerical Methods Applied to Chemical Engineering
Background image of page 2
Image of page 3
This is the end of the preview. Sign up to access the rest of the document.

Page1 / 5

lec16_10182006 - 10.34, Numerical Methods Applied to...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online