An improved linear approximation algorithm for the network equilibrium (packet switching) problem

We present a method for accelerating the convergence of the algorithm based on the Frank and Wolfe linear approximation method for the convex cost multicommodity flow problem known as the "equilibrium traffic assignment problem" in transportation networks and as the "optimal routing o...

Full description

Saved in:
Bibliographic Details
Published in:1977 IEEE Conference on Decision and Control including the 16th Symposium on Adaptive Processes and A Special Symposium on Fuzzy Set Theory and Applications pp. 812 - 818
Main Author: Florian, Michael
Format: Conference Proceeding
Language:English
Published: IEEE 01.12.1977
Subjects:
Acceleration
Approximation algorithms
Communication switching
Convergence
Cost function
Linear approximation
Packet switching
Routing
Telecommunication traffic
Transportation
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We present a method for accelerating the convergence of the algorithm based on the Frank and Wolfe linear approximation method for the convex cost multicommodity flow problem known as the "equilibrium traffic assignment problem" in transportation networks and as the "optimal routing of packet switched messages" in communication networks. The acceleration of the convergence of this algorithm is achieved with a non trivial adaptation of Wolfe's suggestion of an "away" step in the linear approximation method and a variant of this adaptation based on restriction.
DOI:10.1109/CDC.1977.271681