The huddle: combining ai methods to improve handling a group of characters by a human game player
Date
2012
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
The huddle is a concept for extending computer games in which the player controls more than one character. It combines several AI methods to allow the player to create a cooperative strategy for his team of characters. It takes away the need for the player to frantically switch his control from character to character to carry out the strategy that he has in mind. The player enters the huddle from a saved game state where he provides the key ideas of a strategy in the form of situations that his characters can be in and actions for those situations (SAPs). Based upon these ideas, a learner fills in the missing details by adding additional SAPs to create a complete strategy that can be used in the real game. The learning component uses a simulation of the real game, with models of the nonplayer characters created from the observed behaviours of those characters, in order to evaluate candidate solutions. When a solution is found, the player is given the opportunity to take it to the real game and have the behaviour of his characters replaced by the combined strategy created by himself and the huddle. The huddle was evaluated with a fantasy-themed role playing game. It was shown that the huddle allowed the player to focus on developing a strategy without taking away the requirement for him to create the key ideas for that strategy.
Description
Bibliography: p. 97-101
Most pages are in colour.
Most pages are in colour.
Keywords
Citation
Davison, T. (2012). The huddle: combining ai methods to improve handling a group of characters by a human game player (Master's thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca. doi:10.11575/PRISM/4975