This preview shows pages 1–6. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full DocumentThis preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Advanced Mathematical Programming IE417 Lecture 13 Dr. Ted Ralphs IE417 Lecture 13 1 Reading for This Lecture • Chapter 7 IE417 Lecture 13 2 Iterative Algorithms • In previous courses, we have discussed algorithms that were guaranteed to terminate in a finite number of steps, usually with an optimal solution. • For nonlinear optimization, things are not so nice. • We will be dealing with iterative algorithms that produce an infinite sequence of points. • These algorithms may or may not converge to the optimal solution. IE417 Lecture 13 3 Properties of Iterative Algorithms • We will be interested in the following properties of an algorithm: – Does the algorithm converge? * Under what conditions does the algorithm converge? * Does it converge to a global optimal solution? * Does it converge to a local optimal solution? – How quickly does it converge? – How much computational effort is involved in each iteration? – How robust is the algorithm? • We will also be interested in the termination criteria and the accuracy of the solution. IE417 Lecture 13 4 The Algorithmic Map • An algorithm is defined by its algorithmic map . • Given our current location, where do we go next? • This is determined by a mapping A : X → 2 X which maps each point in the domain X to a set of possible “next iterates.” • In other words, if the current iterate is x k , then x k +1 ∈ A ( x k ) ....
View
Full
Document
This note was uploaded on 08/06/2008 for the course IE 417 taught by Professor Linderoth during the Fall '08 term at Lehigh University .
 Fall '08
 Linderoth

Click to edit the document details