Abstract

We wish to explore the contribution that asocial and social learning mightplay as a mechanism for self-adaptation in the search for variable-lengthstructures by an evolutionary algorithm. An extremely challenging, yet simpleto understand problem landscape is adopted where the probability of randomlyfinding a solution is approximately one in a trillion. A number of learningmechanisms operating on variable-length structures are implemented and theirperformance analysed. The social learning setup, which combines forms of bothsocial and asocial learning in combination with evolution is found to be mostperformant, while the setups exclusively adopting evolution are incapable offinding solutions.

 

Quick Read (beta)