Neuroevolution of agents capable of reactive and deliberative behaviours in novel and dynamic environments

Channon

Abstract

Both reactive and deliberative qualities are essential for a good action selection mechanism. We present a model that embodies a hybrid of two very different neural network architectures inside an animat: one that controls their high level deliberative behaviours, such as the selection of sub-goals, and one that provides reactive and navigational capabilities. Animats using this model are evolved in novel and dynamic environments, on complex tasks requiring deliberative behaviours: tasks that cannot be solved by reactive mechanisms alone and which would traditionally have their solutions formulated in terms of search-based planning. Significantly, no a priori information is given to the animats, making explicit forward search through state transitions impossible. The complexity of the problem means that animats must first learn to solve sub-goals without receiving any reward. Animats are shown increasingly complex versions of the task, with the results demonstrating, for the first time, incremental neuro-evolutionary learning on such tasks.

Citation

Channon. (2007). Neuroevolution of agents capable of reactive and deliberative behaviours in novel and dynamic environments. Lecture notes in computer science, 345 - 354. https://doi.org/10.1007/978-3-540-74913-4_35

• Publication Date: Jan 1, 2007
• Journal: Advances in Artificial Life, Proceedings
• Print ISSN: 0302-9743
• Publisher: Springer Verlag
• Pages: 345 - 354
• ISBN: 978-3-540-74912-7
• DOI: https://doi.org/10.1007/978-3-540-74913-4_35
• Keywords: artificial life; neural networks; incremental evolution; reactive and deliberative systems; novel and dynamic environments
• Publisher URL: https://doi.org/10.1007/978-3-540-74913-4_35