Implementation of neural network in game engine to create smart bot and behavior analysis

Abstract

Virtual game players always had a desire for playing with an opponent that acts intelligently like a human. That is why MMO (Massive Multiplayer Online) games have gained huge popularity. Programmers have developed and implemented many systems and algorithms overtime, none came out as successful as Neural Network AI. Artificial Neural Networks (ANN) are computing systems inspired by the biological neural networks that constitute animal brains. Such systems learn (progressively improve performance) to do tasks by considering examples, generally without task-specific programming. We mainly implemented Genetic Algorithm, perceptron algorithm and finally neural network with the help of tensorflow in unity game engine. We finally selected neural network for output efficiency. In this paper, our main focus is to analyze the behavior of the Bots to observe the percentage of efficiency achieved, after the implementation of ANN algorithms in game engine and pointing out the evolving behavior properties. Results from the analysis of our findings can also be helpful for automation and AI development while the whole world is running for these. Neural network algorithms are very complex and a quite time inefficient for video games. But, we implemented ANN in game engine to create smart bots, increasing the efficiency of that algorithm will be another challenge for us. With the help of tensor-flow we made the training process easier, thus making ANN easier to implement in common games.