Social Grammatical Evolution with imitation learning for real-valued function estimation

Drawing on a rich literature concerning social learning in animals, this paper presents a variation of Grammatical Evolution (GE) which incorporates one of the most powerful forms of social learning, namely imitation learning. This replaces the traditional method of `communication' between individuals in GE - crossover - which is drawn from an evolutionary metaphor. The paper provides an introduction to social learning, describes the proposed variant of GE, and tests on a series of benchmark symbolic regression problems. The results obtained are encouraging, being very competitive when compared with canonical GE. It is noted that the literature on social learning provides a number of useful meta-frameworks which can be used in the design of new search algorithms and to allow us to better understand the strengths and weaknesses of existing algorithms. Future work is indicated in this area.

I. Introduction

Many natural computing inspired optimisation algorithms adopt a populational perspective whereby a population of agents search for ‘good’ solutions on a problem landscape [1]. These agents periodically exchange information in order to bias the subsequent search effort, based on feedback concerning the success of the search process to date. Hence, the population ‘learns’ and this leads to adaptation of the search process. While learning is an intuitive concept, it is surprisingly difficult to define comprehensively. Learning can occur on multiple levels, encompassing genetic learning as genes encoding adaptive capabilities in organisms spread through a population over multiple generations, lifetime learning by an individual where the past experiences of an animal impact on its neural wiring and its immune system, and cultural learning such as occurs with humans where useful information is passed from one generation to the next by teaching. Another taxonomy of learning is to divide it into asocial learning, where an organism learns through its own experience, and social learning where an organism learns through interactions with others. Information arising from learning can be distinguished between private and social with the former being personal to an individual, arising from information which the individual has acquired during their lifetime. In contrast, social information is public.