Optimization: principles and algorithms, by Michel Bierlaire
Functions
lineSearch.m File Reference

Algorithm 11.5: Line search algorithm. More...

Go to the source code of this file.

Functions

function lineSearch (in obj, in x, in d, in alpha0, in beta1, in beta2, in lambda, in printlevel)
 Applies the line search algorithm to find a step along a direction that verifies the Wolfe conditions. More...
 

Detailed Description

Algorithm 11.5: Line search algorithm.

Implementation of algorithm 11.5 of [1]

Author
Michel Bierlaire
Date
Fri Mar 20 17:42:10 2015

Definition in file lineSearch.m.

Function Documentation

function lineSearch ( in  obj,
in  x,
in  d,
in  alpha0,
in  beta1,
in  beta2,
in  lambda,
in  printlevel 
)

Applies the line search algorithm to find a step along a direction that verifies the Wolfe conditions.

Note
Tested with runInexactLineSearch.m
Called by steepestDescent
Called by newtonLineSearch
Parameters
objobjective function $f:\mathbb{R}^n \to \mathbb{R} $.
xcurrent iterate
ddirection for the line search
alpha0initial step
beta1parameter for the first Wolfe condition (strictly between 0 and 1. Suggested value: 1.0e-4)
beta2parameter for the second Wolfe condition (strictly between beta1 and 1. Suggested value: 0.99)
lambdaexpansion factor for short steps
printlevelif different from 0, informations are printed at each iteration (Default: 0)
Returns
step alpha verifying the two Wolfe conditions
Copyright 2015-2016 Michel Bierlaire