Intro to Genetic Programming

I found a short paper that has a decent introduction to the concepts of Genetic Programming.

More evolutionary computing

I thought this week I would continue in the same general topic I started last week - evolutionary computing. This week I'd like to talk a little about genetic programming, how it differs from genetic algorithms, and then look at the paper Evolving Computer Programs using Rapidly Reconfigurable Field-Programmable Gate Arrays and Genetic Programming. Don't worry - the length of the name isn't indicative of the length of the paper. This isn't a great intro to GP, but it gives you a feel for a part of evolutionary computing called evolvable hardware. And it also ties in with some other things we've talked about; they discuss optimizing the construction of the sorting networks that Anthony talked about last week.

Routing Protocols for Wireless Ad-hoc Networks

Ad-hoc networks are self organising networks requiring no pre-established infrastructure. Lots of research and efforts have been put into designing an effective routing protocol to effectively determine a route from source to destination. Various table driven, source initiated and position based routing protocol have been proposed. After a quick introduction to ad-hoc networks, if not everyone is familiar with it, i will discuss:

first paper focusing on Review of current ad-hoc routing protocols covering Source initiated Dynamic Source routing (DSR) and table driven Destination Sequenced Distance Vector Routing (DSDV) protocols.

Second paper discussing Position based routing which is a different approach to design an effective routing protocol where location of the hosts (nodes) can be determined by the GPS (Global Position System) or similar methods. I will discuss this approach called as Distance Routing Effect for Mobility (DREAM)


If you are more interested in this topic, and would like to read more. Here are the links:
Survey for various position based ad-hoc routing protocol
Dynamic Source routing (DSR)
Destination Sequenced Distance Vector Routing (DSDV)


--Gaurav

Optimization problems in VLSI circuits

Next Monday, I'll be talking about cell placement in VLSI circuits. I'll do a quick overview of what the problem is and some techniques for performing cell placement, then I will talk about the following paper: Genetic Algorithm for Cell Placement. As part of this, I can give a quick overview of genetic algorithms, if not everyone is familiar with them.

If you are interested in this topic and want a little light reading, this paper has a lot of detail about a number of different techniques, including simulated annealing and genetic algorithms, used for cell placement: VLSI Cell Placement.


Later,
Bob

About Parallel S-trees and POOMA

The powerpoint slides are up on my CS homepage. Here is the Link

I have written answers to some of the questions and listed some relevant URLs. So please take a look at them.

last but not least, I'll be talking about parallel sorting algorithms. while there might be papers out there that could give you an introduction, I think that sorting is pretty self-explanatory, so just come as you are.


anthony

Parallelizing S-trees

Today [April 11], I will be talking about the parallel S-trees.

After a quick introduction to S-trees, i'll present the parallel build and search techniques along with some test results. Most of the content is from the paper Performance Evaluation of Parallel S-trees

-Alok

Distributed is the new Parallel

I'll be stretching the definition of parallel computation for Monday the 11th's class by discussing two papers covering distributed computing.

First, I'll cover the Folding@Home and Genome@Home projects from a high level, looking at what they do and the changes that were required to their algorithms in order to utilize distributed computing. Here's the paper.

Second I'll offer the complement to Chris' discussion of the Google MapReduce library by looking at the Google File System (GFS). I'm planning to offer a fairly detailed look at how those smart kids at Google implemented GFS, keeping in mind that I need to highlight parallelism. Here's the paper.

I've had fun reading about these two rock stars of the distributed computing milieu. Here's hoping that you'll enjoy hearing about them.

Vector Processing with AltiVec and MMX

This (see link below) is a recent article from IBM about AltiVec vector processing support in PowerPC chips in comparison to AMD and Intel chips, and is relevant with respect to the question about the MMX instruction set that came up during my presentation.

AltiVec

Google File System (GFS)

Since a few people were asking about the Google File System, I thought I'd put a link to the reference below...

- Chris

Google File System

4/4/2005 Presentation Slides

I have uploaded my presentation slides for the Google MapReduce and MPI if anyone is interested. Here are the links...

Google MapReduce
Intro to MPI


- Chris