Learning Mental Models

The process of learning is not always as simple as mapping inputs to the best outputs. Often internal state is needed to distinguish between observably identical states of the world. Genetic programming has concentrated on solving problems in the functional/reactive arena, in part because of the absence of a natural way to incorporate memory into the paradigm. This paper presents a simple addition to the genetic programming paradigm that seamlessly incorporates the evolution of the effective gathering, storage, and retrieval of arbitrarily complicated state information. Experimental results show that the effective production and use of complex state structures can be evolved and that agents evolving the use of memory quickly and permanently displace purely reactive and non-deterministic functions. These results may not only aid future research into the causes and constituents of mental models but may expand the types of problems that can be practically tackled by genetic programming.