When Charles Darwin first expounded on his Theory of Evolution in 1859, a computer was an individual who did complex calculations.

Almost 150 years later, a software company is drawing on Darwinian theory to drastically improve the way modern computers solve problems.

Staff at Brighton-based Codefarm are expert in the field of "evolutionary computing", until now largely confined to university research departments.

The technology "educates" computers to solve problems by trial and error as part of a self-regenerating process which eventually provides the best solution.

Codefarm, founded in 2002, is the leader in the field and has already persuaded a number of international investment banks to use its software.

Banks use Codefarm's technology to come up with the best stock and share options for investors who may have given them a complex set of criteria to follow.

Manually it can take anything up to two weeks to compile a portfolio which best suits the investor and which will maximise returns for the bank.

Codefarm's software can churn out portfolios in half an hour, matching, if not improving, the results of number-crunching investment analysts.

Codefarm, in North Road, Brighton, was launched by Sussex University graduates Jeremy Mabbitt, Miles Sabin and Dan Cowan and friend Tony Poppleton.

It moved into the Sussex Innovation Centre at Falmer in October 2003 before moving to bigger headquarters last year to accommodate expansion. It now employs 11 people.

Jeremy said: "The company was founded in order to commercialise evolutionary computing and take it out of the field of research.

"It is something which has always been talked about as a nice idea in theory. We were confident it could be put to good practical use as well.

"People were very sceptical at first because we were presenting them with something that could solve problems they had never identified as problems.

"With the banks we had to demonstrate how it worked by literally timing the software against one of their employees. We set them off at the same time and saw who won.

"Once they can see it works and that it actually improves their returns, they realise it is something they cannot really do without."

Jeremy envisages other applications for evolutionary computing. Travelling salesmen visiting dozens of towns and cities could potentially use the software to work out the best route or telecom companies could one day use it to route calls.

June 3, 2005