MAN – MACHINE INTERFACE

Agents trained by learning techniques provide a powerful approximation of state spaces in games that aretoo large for naive approaches. In the study Genetic Algorithms and Manual Interface was implementedand used to train agents for the board game LUDO. The state space of LUDO is generalized to a small setand encoded to suit the different techniques. The impact of variables and tactics applied in training aredetermined. Agents based on the techniques performed satisfactory against a baseline finite agent, and aGenetic Algorithm based agent performed satisfactory against competitors from the course. Better statespace representations will improve the success of learning based agents.