EnHiC: An Enforced Hill Climbing Based System for General Game Playing

Accurate decision making in games has always been a very complex and yet interesting problem in Artificial Intelligence (AI). General video game playing (GVGP) is a new branch of AI whose target is to design agents that are able to win in every unknown game environment by choosing wise decisions. This paper proposes a new search methodology based on enforced hill climbing for using in GVGP and we evaluate its performance on the benchmarks of the general video game AI competition (GVG-AI). Also a simple and efficient heuristic function for GVGP is proposed. The results show that EnHiC outperforms several well-known and successful methods in the GVG-AI competition.