Optimization: principles and algorithms, by Michel Bierlaire
runRosenbrockSr1.m
Go to the documentation of this file.
1 %> \file
2 %> Runs Trust region method with symmetric rank one (Algorithm 13.2) [\ref newtonTrustRegion] on the Rosenbrock problem.
3 %>
4 %> \note Calls \ref symmetricRankOne
5 %>
6 %> @ingroup Running
7 %> @ingroup chap13
8 %> @author <a href="http://people.epfl.ch/michel.bierlaire">Michel Bierlaire</a>
9 %> @date Sat Mar 21 16:46:45 2015
10
11 %>
12
13 delta0 = 10 ;
14 printf("Dogleg\n") ;
15 x0 = [-1.5 ; 1.5] ;
16 [solution, iteres, niter] = symmetricRankOne('exRosenbrock',x0,delta0,1.0e-8,0,1000) ;
17 format short e
18 iteres(1:niter,:)
19 solution
20 format
21 niter
22
23 printf("Truncated trust region\n") ;
24 x0 = [-1.5 ; 1.5] ;
25 [solution, iteres, niter] = symmetricRankOne('exRosenbrock',x0,delta0,1.0e-8,1,1000) ;
26 format short e
27 iteres(1:niter,:)
28 solution
29 format
30 niter
31
32
function symmetricRankOne(in obj, in x0, in delta0, in eps, in tr)
Applies SR1 algorithm with trust region to solve where . The parameters of the method are taken from...
function exRosenbrock(in x)
Example presented in Section 11.6 of .
function newtonTrustRegion(in obj, in x0, in delta0, in eps, in tr)
Applies Newton&#39;s algorithm with trust region to solve where . The parameters of the method are taken...