Presenter Information

Kira ToalFollow

Start Date

16-12-2020 9:35 AM

Description

For many video game developers, designing an adversary or adversarial environment that is engaging but not overly challenging for a variety of players with different skill levels can be a difficult task. As a result, difficulty adjustment (or difficulty balancing), the process of adjusting game parameters and adversarial policies in order to adapt to a player's skill level and or play-style, has increased in popularity in the worlds of industry and academia alike. Many games that employ difficulty adjustment algorithms, however, are limited in that they rely on player feedback, they can only adjust adversarial policies once the player has reached a certain checkpoint, or they are only capable of changing one game parameter at a time. This research seeks to improve upon existing methods of difficulty balancing by creating an artificially intelligent agent that utilizes the Minimax Algorithm to dynamically adjust the level design of a 2D platformer-style video game in accordance with player acclimation to in-game adversarial policies.

Comments

Mentor: Andrew Forney

Share

COinS
 
Dec 16th, 9:35 AM

Building a Smarter Video Game Adversary With Dynamic Difficulty Adjustment

For many video game developers, designing an adversary or adversarial environment that is engaging but not overly challenging for a variety of players with different skill levels can be a difficult task. As a result, difficulty adjustment (or difficulty balancing), the process of adjusting game parameters and adversarial policies in order to adapt to a player's skill level and or play-style, has increased in popularity in the worlds of industry and academia alike. Many games that employ difficulty adjustment algorithms, however, are limited in that they rely on player feedback, they can only adjust adversarial policies once the player has reached a certain checkpoint, or they are only capable of changing one game parameter at a time. This research seeks to improve upon existing methods of difficulty balancing by creating an artificially intelligent agent that utilizes the Minimax Algorithm to dynamically adjust the level design of a 2D platformer-style video game in accordance with player acclimation to in-game adversarial policies.