Fractal Sets Records With One Million Antenna Simulations


BOSTON--(BUSINESS WIRE)--Aug. 13, 2002--Today, Fractal Antenna Systems, Inc., announced that its Genetic Optimizer Cluster (GO Cluster) set a new record in its ability to simulate and solve problems in antenna design and electronics. With over 1,000,000 antenna simulations created and assessed, the firm's latest update of its innovative tool has already demonstrated its worth in new and advantageous antenna and circuit designs in just a few weeks of computational `turn on'.

Using a proprietary genetic algorithm, which mimics evolution's natural selection, the firm took two big steps to speed up and increase the ability of this optimization method to solve complex problems. Based on 1997 work by the Fractal Antenna Systems' Chief Technical Officer, Dr. Nathan Cohen, the GO Cluster uses a novel way of detailing and compressing the parameters of a problem, using a `fractal coding' of the genome. The patent pending method and apparatus allows the opportunity for far faster convergence in most optimization problems. Next, the running of many simulations simultaneously was made possible by hooking up 16 or more PCs in a computer cluster. Churning away, often at 25 or more simulations per minute, the GO Cluster can hunt and seek the best design structures for antennas that need to be better than conventional designs.

Cohen notes: "Design space afforded by fractal geometry is virtually endless. This gives a wonderful range of options for solutions but makes an exhaustive unguided search intractable. In comparison, the tiny set of conventional geometric designs are ill-equipped to attack complex electromagnetic problems such as posed by modern antenna design. As someone who has conceived, built and measured a few thousand fractal antennas in 14 years, I am relieved that we now have this wonderful proprietary tool. Its far easier and faster to find the best fit to the problem under attack with the GO Cluster."

The more than a million antenna design simulations are closely approximating the total number of publicly known antenna simulations done by researchers in the last few years. The GO Cluster is expected to exceed the total number of all such antenna simulations done up to now in the coming year. "It's a treasure hunt--the hunt for the better antenna--and we have the tool to dig it out," notes Cohen.

Cohen also notes that the GO Cluster has just solved a key problem in antenna arrays, resolving an issue that has been debated for many years. "There is great science and engineering in the GO cluster, and simulating many cases in the fractal design space is the only way to go; this is where the frontier of electromagnetics and antenna engineering lies."
 

 CONTACT: FAS