Multi-objective topology and weight evolution of neuro-controllers

Omer Abramovich; Amiram Moshaiov

doi:10.1109/CEC.2016.7743857

Multi-objective topology and weight evolution of neuro-controllers

Omer Abramovich, Amiram Moshaiov

School of Mechanical Engineering

Tel Aviv University

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

18 Scopus citations

Abstract

Evolutionary multi-objective optimization has been employed in studies concerning evolutionary robotics, and in particular for the evolution of neuro-controllers. To allow the simultaneous multi-objective evolution of topology and weights, tailored search algorithms should be developed. Here, a modification to the well-known NEAT algorithm is suggested. The proposed algorithm, which is termed NEAT-MODS, involves a specialized selection process that aims to ensure both genotypic diversity and elitism in the context of Pareto-optimality. NEAT-MODS constitutes a generic Multi-objective Topology and Weight Evolution of Artificial Neural-Networks (MO-TWEANN) algorithm. The suggested NEAT-MODS is found to be statistically superior to NEAT-PS, when applied to solve complex multi-objective navigation problem.

Original language	English
Title of host publication	2016 IEEE Congress on Evolutionary Computation, CEC 2016
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	670-677
Number of pages	8
ISBN (Electronic)	9781509006229
DOIs	https://doi.org/10.1109/CEC.2016.7743857
State	Published - 14 Nov 2016
Event	2016 IEEE Congress on Evolutionary Computation, CEC 2016 - Vancouver, Canada Duration: 24 Jul 2016 → 29 Jul 2016

Publication series

Name	2016 IEEE Congress on Evolutionary Computation, CEC 2016

Conference

Conference	2016 IEEE Congress on Evolutionary Computation, CEC 2016
Country/Territory	Canada
City	Vancouver
Period	24/07/16 → 29/07/16

Access to Document

10.1109/CEC.2016.7743857

Cite this

@inproceedings{be5e9799a942411fb6b8ed125c635b1d,

title = "Multi-objective topology and weight evolution of neuro-controllers",

abstract = "Evolutionary multi-objective optimization has been employed in studies concerning evolutionary robotics, and in particular for the evolution of neuro-controllers. To allow the simultaneous multi-objective evolution of topology and weights, tailored search algorithms should be developed. Here, a modification to the well-known NEAT algorithm is suggested. The proposed algorithm, which is termed NEAT-MODS, involves a specialized selection process that aims to ensure both genotypic diversity and elitism in the context of Pareto-optimality. NEAT-MODS constitutes a generic Multi-objective Topology and Weight Evolution of Artificial Neural-Networks (MO-TWEANN) algorithm. The suggested NEAT-MODS is found to be statistically superior to NEAT-PS, when applied to solve complex multi-objective navigation problem.",

author = "Omer Abramovich and Amiram Moshaiov",

note = "Publisher Copyright: {\textcopyright} 2016 IEEE.; 2016 IEEE Congress on Evolutionary Computation, CEC 2016 ; Conference date: 24-07-2016 Through 29-07-2016",

year = "2016",

month = nov,

day = "14",

doi = "10.1109/CEC.2016.7743857",

language = "אנגלית",

series = "2016 IEEE Congress on Evolutionary Computation, CEC 2016",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "670--677",

booktitle = "2016 IEEE Congress on Evolutionary Computation, CEC 2016",

address = "ארצות הברית",

}

Abramovich, O & Moshaiov, A 2016, Multi-objective topology and weight evolution of neuro-controllers. in 2016 IEEE Congress on Evolutionary Computation, CEC 2016., 7743857, 2016 IEEE Congress on Evolutionary Computation, CEC 2016, Institute of Electrical and Electronics Engineers Inc., pp. 670-677, 2016 IEEE Congress on Evolutionary Computation, CEC 2016, Vancouver, Canada, 24/07/16. https://doi.org/10.1109/CEC.2016.7743857

Multi-objective topology and weight evolution of neuro-controllers. / Abramovich, Omer; Moshaiov, Amiram.
2016 IEEE Congress on Evolutionary Computation, CEC 2016. Institute of Electrical and Electronics Engineers Inc., 2016. p. 670-677 7743857 (2016 IEEE Congress on Evolutionary Computation, CEC 2016).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › peer-review

TY - GEN

T1 - Multi-objective topology and weight evolution of neuro-controllers

AU - Abramovich, Omer

AU - Moshaiov, Amiram

PY - 2016/11/14

Y1 - 2016/11/14

N2 - Evolutionary multi-objective optimization has been employed in studies concerning evolutionary robotics, and in particular for the evolution of neuro-controllers. To allow the simultaneous multi-objective evolution of topology and weights, tailored search algorithms should be developed. Here, a modification to the well-known NEAT algorithm is suggested. The proposed algorithm, which is termed NEAT-MODS, involves a specialized selection process that aims to ensure both genotypic diversity and elitism in the context of Pareto-optimality. NEAT-MODS constitutes a generic Multi-objective Topology and Weight Evolution of Artificial Neural-Networks (MO-TWEANN) algorithm. The suggested NEAT-MODS is found to be statistically superior to NEAT-PS, when applied to solve complex multi-objective navigation problem.

AB - Evolutionary multi-objective optimization has been employed in studies concerning evolutionary robotics, and in particular for the evolution of neuro-controllers. To allow the simultaneous multi-objective evolution of topology and weights, tailored search algorithms should be developed. Here, a modification to the well-known NEAT algorithm is suggested. The proposed algorithm, which is termed NEAT-MODS, involves a specialized selection process that aims to ensure both genotypic diversity and elitism in the context of Pareto-optimality. NEAT-MODS constitutes a generic Multi-objective Topology and Weight Evolution of Artificial Neural-Networks (MO-TWEANN) algorithm. The suggested NEAT-MODS is found to be statistically superior to NEAT-PS, when applied to solve complex multi-objective navigation problem.

UR - http://www.scopus.com/inward/record.url?scp=85008255723&partnerID=8YFLogxK

U2 - 10.1109/CEC.2016.7743857

DO - 10.1109/CEC.2016.7743857

M3 - ???researchoutput.researchoutputtypes.contributiontobookanthology.conference???

AN - SCOPUS:85008255723

T3 - 2016 IEEE Congress on Evolutionary Computation, CEC 2016

SP - 670

EP - 677

BT - 2016 IEEE Congress on Evolutionary Computation, CEC 2016

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2016 IEEE Congress on Evolutionary Computation, CEC 2016

Y2 - 24 July 2016 through 29 July 2016

ER -

Multi-objective topology and weight evolution of neuro-controllers

Abstract

Publication series

Conference

Access to Document

Other files and links

Fingerprint

Cite this