Author: pelikan

The simulated car racing competition organized at GECCO-2009 is part of the 2009 Simulated Car Racing Championship, an event joining the three competitions held at CEC-2009, GECCO-2009, and CIG-2009. Pier Luca Lanzi reports on the first contest in this competition (Contest 1: Learning to Drive) here. The deadline for the submissions to this contest is July 1, 2009.

Click here for another related post at IlliGAL Blog. The photo is due to SPYDERMAN360 (wikipedia).

It’s now less than a month until the deadline for paper submission for GECCO-2009, which will be the first GECCO conference in Canada. The conference will take place in Montreal, Canada on July 8-12, 2009 (Wednesday-Sunday). The deadline for submitting papers to GECCO-2009 is January 14, 2009. The general chair of GECCO-2009 is Franz Rothlauf.

Besides the 40 tutorials, 13 workshops, and several great competitions, GECCO-2009 will feature a number of other events, including student party, social event with dinner and birthday cake, poster session and hors d’oeuvre reception, art exhibition, and 3rd Annual Job Shop.

For more details, check out GECCO-2009 web site. Download the PDF of the GECCO-2009 call for papers here. GECCO banners can be found here.

Marcelo De Brito of Genetic Argonaut pointed out an interesting article published by wired.com on top national security challenges for the next president.

One of these challenges is the use of genetic algorithms for battlefield operations. This reminded me of the project FOX-GA, which was one of the projects I’ve learned about while I was at the Illinois Genetic Algorithms Laboratory.

About a year ago I started a project with Michal Hammel from the Lawrence Berkeley National Lab on using a genetic algorithm for modeling flexible macromolecular systems (more specifically, the focus was on large proteins of over 900 amino acids).

To make the long story short, some proteins do not have rigid structure, but their structure changes over time. These proteins would typically contain several rigid modules, which are connected with flexible linkers. The goal of this project is to find out how the structure changes over time based on experimental results from solution-based SAXS (small angle X-ray scattering) and theoretical conformations computed with molecular dynamics (MD) simulations.

The basic idea of this approach called BILBOMD follows:
1. Compute experimental scattering profile using solution-based SAXS.
2. Compute a large number (1k to 10k) of potential conformations using molecular dynamics simulation.
3. Use a genetic algorithm to select a subset of conformations explaining the experimental scattering profile best.

The results are promising and we’re working on making these results even better and more useful.

A web page dedicated to this project can be found here.