Advanced search
Start date
Betweenand

Automatic evolution of behaviour trees for an intelligent agent

Grant number: 13/20606-0
Support type:Scholarships in Brazil - Scientific Initiation
Effective date (Start): November 01, 2013
Effective date (End): November 30, 2015
Field of knowledge:Physical Sciences and Mathematics - Computer Science - Computer Systems
Principal Investigator:Vinícius Veloso de Melo
Grantee:Léo Françoso Dal Piccol Sotto
Home Institution: Instituto de Ciência e Tecnologia (ICT). Universidade Federal de São Paulo (UNIFESP). Campus São José dos Campos. São José dos Campos , SP, Brazil

Abstract

This project aims to study the application of evolutionary algorithms - Genetic Programming initially - in the automatic evolution of behaviour trees of an intelligent agent, which determine the actions to be taken in accordance with the current configuration of the environment. The intelligent agent that will be studied in this project is an implementation in the framework for study of artificial intelligence tools called "The AI Sandbox" (TAS). This agent is the commander of a small military squad that acts against another squad in a security exercise known as "Capture The Flag" (CTF), in which the goal is to capture and bring the enemy flag to the Allied camp. For the intelligent agent can accomplish this task, a behaviour tree will be evolved based on a behaviour tree of an existing commander in the tool. One aims to assess the ability of the Genetic Programming technique to evolve one or more behaviour trees that outperform the original one.

Scientific publications
(References retrieved automatically from Web of Science and SciELO through information on FAPESP grants and their corresponding numbers as mentioned in the publications by the authors)
DAL PICCOL SOTTO, LEO FRANCOSO; DE MELO, VINICIUS VELOSO; BASGALUPP, MARCIO PORTO. -LGP: an improved version of linear genetic programming evaluated in the Ant Trail problem. KNOWLEDGE AND INFORMATION SYSTEMS, v. 52, n. 2, p. 445-465, AUG 2017. Web of Science Citations: 2.
DAL PICCOL SOTTO, LEO FRANCOSO; DE MELO, VINICIUS VELOSO. Studying bloat control and maintenance of effective code in linear genetic programming for symbolic regression. Neurocomputing, v. 180, n. SI, p. 79-93, MAR 5 2016. Web of Science Citations: 5.

Please report errors in scientific publications list by writing to: cdi@fapesp.br.