An Agent Inspired Reconfigurable Computing Implementation of a Genetic Algorithm

emerge from when multiple agents interact with one another. Many software systems have been successfully imAlthough many competing definitions of what conplemented using an agent paradigm which employs stitutes an agent exist there is generally good agreea number of independent entities that communicate ment that agents should operate in a system with with one another to achieve a common goal. The other agents and that the agents should display distributed nature of such a paradigm makes it an the characteristics of autonomy and sociability [I]. excellent candidate for use in high speed reconfigAgents are autonomous in that they operate indeurable computing hardware environments such as pendently and asynchronously. Each agent makes its those present in modem FPGA’s. In this paper, a own decisions on actions to take and does not operdistributed genetic algorithm that can be applied to ate either as a master or a slave. Agents are social in the agent based reconfigurable hardware model is inthat they communicate with other agents in the systroduced. The effectiveness of this new algorithm is tem to exchange data. Multiagent systems are comevaluated by comparing the quality of the solutions prised of potentially many independent, autonomous, found by the new algorithm with those found by trarelatively simple agents that communicate with one ditional genetic algorithms. The performance of a another and cooperate and/or compe to achieve a reconfigurable hardware implementation of the new common goal. algorithm on an FPGA is compared to traditional single processor implementations. 2 Reconfigurable Computing Implementation of Agents