Dynamic difficulty adjustment of game AI for video game Dead-End

To create a satisfactory game opponent is to optimize player's experience through creation of an dynamic balanced game, which means that win-rate of players is adjusted according to their ability. The most commonly used approach for generating satisfactory game opponent is Dynamic Difficulty Ad...

Full description

Saved in:
Bibliographic Details
Main Authors: Xinrui Yu, Suoju He, Yuan Gao, Jiajian Yang, Lingdao Sha, Yidan Zhang, Zhaobo Ai
Format: Conference Proceeding
Language:English
Subjects:
ANN
Artificial intelligence
Artificial neural networks
clustering
Clustering algorithms
DDA
Dogs
Games
Helium
Humans
player strategy
Software engineering
Testing
Unsupervised learning
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:To create a satisfactory game opponent is to optimize player's experience through creation of an dynamic balanced game, which means that win-rate of players is adjusted according to their ability. The most commonly used approach for generating satisfactory game opponent is Dynamic Difficulty Adjustment (DDA), which is to dynamically adjust challenge level of the opponent according to the player's skill level. However, DDA currently used is relatively simple and implementing DDA by adjusting opponent's intelligence is still challenging. In this paper, we propose to use Artificial Neural Network(ANN) to implement DDA and unsupervised learning methodologies to improve the performance of ANN. ANN-controlled Non-Player Characters (NPC) can make "wise" decision based on collected attributes of all the characters in the game. Different ANNs can provide different win-rates for different player strategies, which can achieve the dynamic balance we expected and enhance the user experience of games.
DOI:10.1109/ICICIS.2010.5534761