An Evolutionary Algorithm Inspired by Evolution of Birdsong Grammars to Quickly Obtain Practical Solutions to Large-scaled Problems

We propose an evolutionary algorithm inspired by evolution of birdsong grammars. The proposed algorithm is designed to quickly obtain practical solutions to large-scaled problems. The main mechanism for obtaining practical solutions quickly is to represent solution candidates in a compact way using grammars, that is to say, to make the search space smaller than the original one using grammars. Also, we apply the algorithm to a large-scaled OneMax problem and its variants. The simulation results show that the algorithm tends to quickly produce solution candidates with simple bit patterns and with not so bad fitness values for large-scaled problems as a bit pattern of the global optimum becomes more random. That can be considered as that practical solutions to large-scaled problems are obtained quickly. It is also shown that when the problem size is smaller, the algorithm tends to produce a precise bit pattern of the global optimum.