InterJournal Complex Systems, 1061
Status: Accepted
Manuscript Number: [1061]
Submission Date: 2004
Using Product Distributions for Distributed Optimization
Author(s): Stefan Bieniawski ,David H. Wolpert

Subject(s): CX.6



With connections to bounded rational game theory, information theory and statistical mechanics, Product Distribution (PD) theory provides a framework for performing distributed optimization. This paper will present the mathematical foundations and the details of the algorithm used to perform optimization. The focus will be on understanding the mechanisms which are achieving the coordination and on highlighting the flexibility of the approach. The technique will be demonstrated on a variety of unconstrained and constrained optimization problems. The talk will also include comparisons with two forms of distributed reinforcement learning inspired optimization approaches, parallel and serial Brouwer updating. The inter-relationship of the techniques will be discussed.

Retrieve Manuscript
Submit referee report/comment

Public Comments: