PROJECT TITLE :

Predicting Opponent's Production in Real-Time Strategy Games With Answer Set Programming

ABSTRACT:

The adversarial character of real-time strategy (RTS) games is one amongst the most sources of uncertainty at intervals this domain. Since players lack actual information regarding their opponent's actions, they need a affordable illustration of different prospects and their probability. In this article we propose a methodology of predicting the most probable combination of units made by the opponent throughout a certain time period. We have a tendency to employ a logic programming paradigm referred to as Answer Set Programming, since its semantics is well fitted to reasoning with uncertainty and incomplete information. In contrast with typical, purely probabilistic approaches, the presented technique takes under consideration the background data regarding the sport and only considers the mixtures that are in line with the sport mechanics and with the player's partial observations. Experiments, conducted throughout different phases of StarCraft: Brood War and Warcraft III: The Frozen Throne games, show that the prediction accuracy for time intervals of 1-three min appears to be surprisingly high, creating the tactic useful in apply. Root-mean-square error grows only slowly with increasing prediction intervals-virtually in an exceedingly linear fashion.