Distributed-Genetic-Algorithm

Used genetic algorithm to significantly reduce the number of feature in the data set without compromising with the classification accuracy. A distributed 1-N master-worker model was adopted. Docker images for the workers were created and then deployed on multiple Amazon web services EC2 instances. Master is the primary controller of the system which makes asynchronous requests to multiple workers running in different containers. Each worker returns the evaluated fitness value in response to the master, where master analyses the response and requests again until the threshold classification value is achieved.