War machines G: a partial-order planning system for a war game domain

• Main Authors: Macaspac, Kane Chelster M., Sta. Ana, Jeffrey B., Wong, Madeleine Ariane A.
• Format: text
• Language: English
• Published: Animo Repository 2004
• Subjects: Computer games > Programming; Computer graphics; Artificial intelligence; Intelligent agents (Computer software); Computer Sciences; Programming Languages and Compilers; Software Engineering
• Online Access: https://animorepository.dlsu.edu.ph/etd_bachelors/6604
• Institution: De La Salle University

With the emergence of the use of AI in computer games, players have been faced with level of game-play and challenge that provide more enjoyment. Although techniques like neural networks and genetic algorithms have been used by some of the more revolutionary games, most AI used in games are based on the more common finite state machines and rule-based systems. The proponents propose the use of AI planning in the use of computer games, specifically one that is based on the partial-order planning algorithm (POP). To concretize the research, the War Machines system was implemented as a war game domain that uses a planning system based on POP to control a team in the game. To accommodate the various aspects of the war game domain, the planning algorithm interleaves planning with execution, makes use of observational actions and implements LCW to prevent redundant sensing. Testing of the system showed the planner to sufficiently attach actions to goals but implementation of some aspects of the system caused problems such as the planner's team to unknowingly achieve the main goal. It has been concluded that to be successful with the use of POP, a good deal of time should be spent in designing the action domain.