Abstract :
In this paper, we propose a new double-layered state space construction method, which consists of Fritzke´s Growing Neural Gas algorithm and a class management mechanism of GNG units. The classification algorithm yields a new class by referring to anticipation error, anticipation vectors of an originated class, and anticipation vectors GNG units belonging in the originated class.
Keywords :
learning (artificial intelligence); multi-agent systems; neural nets; Fritzke Growing Neural Gas algorithm; GNG units; anticipation error; anticipation vectors; class management mechanism; competitive learning neural network; double layered incremental state space construction method; reinforcement learning agents;
