Which is best approach to choose for game playing problem?
The original questions seems to seek a general AI approach to game play, requiring the top advancement of the last few years to solve. The answer Can an AI learn how to play chess without instructions? outlines the current leading edge approach to relatively game-independent learning of champion level game play.
In the comments below the question, the eight queens puzzle is mentioned, which is not like the tournament games. It is a very specific puzzle, too simple to solve with AI, although as an proof of concept for a general problems solver approach or a search prioritization strategy it may suffice.