Tetris Player: Strategy Driven Algorithm
Martin Lukáč, Paul Bourgine · 2000
In this paper we explore the problem of computational artificial life. For this purpose we choose to model Tetris, a simple ruled game. We propose a simple algorithm based on the duality of exploration and exploitation as a base for a powerful prediction system. This is done by the strategy driven algorithm. We use the TD(λ) reinforcement learning in combination with a multilayer perceptron.