An Improved Chess AI System with Hybrid-Fly Optimization and Artificial Neural Networks

The "Drosophila of Artificial Intelligence (AI)" is a chessboard, since chess was an early and popular subject in AI studies, and the field later expanded to include other games. The majority of the most advanced chess engines relied on more traditional search strategies like as Minimax and Alpha beta prior to the current success of systems that are based on notions from machine learning. The majority of chess engines aren't built to mimic human intuition, but rather to compute as many feasible permutations before making a move. New features in the game demonstrated how AlphaZero used a deep reinforcement learning-based strategy. Training the model with top human games may increase chess engines' intuitive sense for the situation, and our suggested hybrid optimization approach can glean the ideal move. This study addresses the shortcomings of traditional methods of evaluating a chess situation and presents a new model, the Hybridfly-ANN, to achieve the greatest number of legal moves. The suggested technique first determines all legitimate moves that might be made in response to a single opponent move, and then uses the standard mini-max algorithm to determine which move is the most optimum. An important part of this study is the suggested Hybrid Fly optimization approach, which improves the system's performance by fine-tuning the ANN classifier's weights. The effectiveness of the method that has been suggested is assessed using a series of evaluation criteria that includes accuracy, sensitivity, and specificity.