Learning Behavior Trees for Autonomous Agents with Hybrid Constraints Evolution

Created by W.Langdon from gp-bibliography.bib Revision:1.7810

@Article{zhang:2018:AS,

• author = "Qi Zhang and Jian Yao and Quanjun Yin and Yabing Zha",
• title = "Learning Behavior Trees for Autonomous Agents with Hybrid Constraints Evolution",
• journal = "Applied Sciences",
• year = "2018",
• volume = "8",
• number = "7",
• keywords = "genetic algorithms, genetic programming",
• ISSN = "2076-3417",
• URL = "https://www.mdpi.com/2076-3417/8/7/1077",
• DOI = "doi:10.3390/app8071077",
• abstract = "In modern training, entertainment and education applications, behaviour trees (BTs) have already become a fantastic alternative to finite state machines (FSMs) in modelling and controlling autonomous agents. However, it is expensive and inefficient to create BTs for various task scenarios manually. Thus, the genetic programming (GP) approach has been devised to evolve BTs automatically but only received limited success. The standard GP approaches to evolve BTs fail to scale up and to provide good solutions, while GP approaches with domain-specific constraints can accelerate learning but need significant knowledge engineering effort. In this paper, we propose a modified approach, named evolving BTs with hybrid constraints (EBT-HC), to improve the evolution of BTs for autonomous agents. We first propose a novel idea of dynamic constraint based on frequent sub-trees mining, which can accelerate evolution by protecting preponderant behaviour sub-trees from undesired crossover. Then we introduce the existing static structural constraint into our dynamic constraint to form the evolving BTs with hybrid constraints. The static structure can constrain expected BT form to reduce the size of the search space, thus the hybrid constraints would lead more efficient learning and find better solutions without the loss of the domain-independence. Preliminary experiments, carried out on the Pac-Man game environment, show that the hybrid EBT-HC outperforms other approaches in facilitating the BT design by achieving better behaviour performance within fewer generations. Moreover, the generated behaviour models by EBT-HC are human readable and easy to be fine-tuned by domain experts.",
• notes = "also known as \cite{app8071077}",

}

Genetic Programming entries for Qi Zhang Jian Yao Quanjun Yin Yabing Zha

Citations