The aim of this paper is to develop and study an artificial intelligence based game-playing agent using genetic algorithm and neural networks. We first create an agent which learns how to optimally play the famous “Flappy Bird” game by safely dodging all the barriers and flapping its way through them and then study the effect of changing various parameters like number of neurons on the hidden layer, gravity, speed, gap between trees has on the learning process. The gameplay was divided into two level of difficulty to facilitate study on the learning process. Phaser Framework was used to facilitate HTML5 programming for introducing real-life factors like gravity, collision and Synaptic Neural Network library was used to implement neural network so as to avoid creating a neural network from scratch. Machine Learning Algorithm which we have adopted in this project is based on the concept of Neuro-Evolution and this form of machine learning uses algorithms which can evolve and mature over time such as a genetic algorithm to train artificial neural networks.
CITATION STYLE
Mishra, Y., Kumawat, V., & Selvakumar, K. (2019). Performance analysis of flappy bird playing agent using neural network and genetic algorithm. In Communications in Computer and Information Science (Vol. 1025 CCIS, pp. 253–265). Springer. https://doi.org/10.1007/978-981-15-1384-8_21
Mendeley helps you to discover research relevant for your work.