Artificial Asymmetric Potential Fields

Potential field is a technique that is used for the navigation of objects in the game environment in order to achieve the natural movements. In this thesis we extended this work by using the asymmetric potential fields in scenarios of one and two types of units. By using asymmetric potential fields we achieved the repelling and attracting behavior of units at the same time. We conducted matches of bots implemented with the symmetric and asymmetric potential fields against the bots who participated in the first tournament of StarCraft competition. We used these bots as a bench mark for the validation of results. We analyzed that the bot with asymmetric potential fields performed better as compared to the bot with symmetric potential fields regarding the completion time, win ratio and number of leader killed. We validated the result by t-test which proves that there was significance difference between the performance of the asymmetric potential field bot and symmetric potential field bot.