Abstract
Artificial intelligence is an important topic for video games. It can be used to simulate human-like behavior for non-player characters (NPCs). NPCs can be used to create opponents in a game or to enhance the universe of a game. Introducing NPCs in a competitive strategy game can help to improve the long time play value of the game. NPCs could, for example, be used to implement in-game events, where players have to ally to defeat an NPC. Additionally, a computer program acting like a human player could be used to perform real-life server load tests to check if changes in, for example, database queries have an effect on the performance of the game, before they go into production mode.
This thesis discusses the conception and development of the first artificial intelligence for the game Grepolis. In this project, different strategies for artificial intelligence in video games, with a focus on decision making, are evaluated how good they to the game. Important questions concerning the implementation of artificial intelligence in video games are asked and answered for this project. A concept for an artificial intelligent NPC for the game Grepolis is built and implemented. Finally an evaluation is done to prove that the implemented solution corresponds to the requirements of the project.
After completing the project, InnoGames has a functional artificial intelligence for the game Grepolis that simulates human-like behavior. The developed solution uses the well known decision making techniques decision trees, state machines and goal oriented behavior to create a new way of decision making.
Information
Space of time: 02/2016 – 09/2016
Coding Language: PHP
Team: alone
Background: Master thesis during my work as a working student at InnoGames
Report: Click here