I can’t believe we are in the double digits discussion activities! This semester is blowing by … its been fun so far. Anyway, you ought to spend a little time reading Topic 10 … it will give you some good insight into the types of systems you’ll need to plan for when solving the 110 Challenge. Ok, here we go …
Computers that can evolve to improve themselves have been fantasized in science fiction books for years, but something similar has occurred at the University of Sussex in England. Adrian Thompson (Center for Computational Neuroscience and Robotics) works with computer chips that can manage their own logic gates by testing new designs and choosing the best configurations for a particular task. Two technologies make this possible: evolutionary algorithms are computer programs that can rapidly generate variations in their own code to evaluate and select the most efficient code; and Field Programmable Gate Arrays (FPGAs), which use transistors that appear as an array of logic cells to change their value and connect to any other logic cells as they are programmed. When these two technologies are brought together (shazam!), circuits can become more effective than similar circuits designed by humans using known principles. There is one problem — Adrian doesn’t know how it works.
So if a management system, which is a critical system, has been built using a working technology such as evolutionary algorithms and FPGAs, but we don’t know how it works, should we implement this technology? Is this technology reliable? Explain the possible advantages and disadvantages, potential for success and potential risk in your opinion.