An Agent Inspired Reconfigurable Computing Implementation of a Genetic Algorithm

摘要

Many software systems have been successfully implemented using an agent paradigm which employs a number of independent entities that communicate with one another to achieve a common goal. The distributed nature of such a paradigm makes it an excellent candidate for use in high speed reconfigurable computing hardware environments such as those present in modem FPGA's. In this paper, a distributed genetic algorithm that can be applied to the agent based reconfigurable hardware model is introduced. The effectiveness of this new algorithm is evaluated by comparing the quality of the solutions found by the new algorithm with those found by traditional genetic algorithms. The performance of a reconfigurable hardware implementation of the new algorithm on an FPGA is compared to traditional single processor implementations.