Implementation Method Of Artificial Intelligence Of Non-Player Character In Role-Playing Game

In recent years,with the rapid development of computer hardware and software level,video game industry developed rapidly.Role-playing games occupies an important market share.In addition to the game audio-visual effects,the game’s playability as a deeper level of the game get more and more attention by the game developers.Non-player character’s artificial intelligence design is conducive to reflect the game’s playability.In the current role-playing game scene,the non-player character’s artificial intelligence model is reflected by a lot of variables,leading to excessive decision-making branch,non-player role artificial intelligence development efficiency is affected.This paper presents a new way of implementing non-player character artificial intelligence,that is behavior tree model combined with machine learning algorithm,which aims to transform the pure behavior tree decision model of non-player character into a decision model combined with behavior tree and classifier.The method of making low-level decision-making model by game developers changes from traditional generating behavior tree to generating classifier,and the workload of generating classifier training data is less than the workload of generating behaviour tree,so it can reduce the workload of game developers.This paper designs and implements three tools: the data generation tool,the Bp_Adaboost model generation tool,and the behavior tree tool that combines Bp_Adaboost.These three tools are mainly used to solve the problem of making the classifier problem and the combination of classifier and behavior tree in the pratical game development.The data generation tool is used to generate classifier training data that is difficult to obtain directly from the game;the Bp_Adaboost model generation tool is used to train the classifier;the Bp_Adaboost behavior tree tool is used to combine the classifier with the behavior tree.Finally,this article uses the Unity3 D game engine to develop a single role-playing game with combat as the main game content.The game is based on this model to realize the artificial intelligence of non-player character.The experimental results show that the model is effective in the development of artificial intelligence of non-player character,and this model has higher development efficiency compared with traditional artificial intelligence model.