Human-like decision making for bots inmobile gaming : A case study in developing believable artificial intelligence for mobile games

Abstract: An AI framework for game AI of bots has been developed together with the company Level Eight for their upcoming game PROTOSTRIKE. The primary goal of the thesis is to implement a decision maker in the framework that exhibits human-like behaviours. An overview and evaluation of existing decision techniques in the games industry have been done, and together with information gathered from the team at Level Eight, Utility AI was selected as most suitable for the decision-making system for the bots. The implemented prototype of the AI framework is evaluated about the bots’ ability to appear human-like in a multi-player environment. The results indicate that the lack of focus is the major behaviour that makes a bot not to act human-like. The test results have also been used to provide suggestions for future work and conclusion about the thesis as a whole. Due to problems during testing, no statistical conclusions can be made. However, the comments made by the test subjects added insight how to improve the AI framework.