Forward-Backward Rapidly-Exploring Random Trees for Stochastic Optimal Control

We propose a numerical method for the computation of the forward-backward stochastic differential equations (FBSDE) appearing in the Feynman-Kac representation of the value function in stochastic optimal control problems. By the use of the Girsanov change of probability measures, it is demonstrated...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Hawkins, Kelsey P, Pakniyat, Ali, Theodorou, Evangelos, Tsiotras, Panagiotis
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
container_end_page
container_issue
container_start_page
container_title
container_volume
creator Hawkins, Kelsey P
Pakniyat, Ali
Theodorou, Evangelos
Tsiotras, Panagiotis
description We propose a numerical method for the computation of the forward-backward stochastic differential equations (FBSDE) appearing in the Feynman-Kac representation of the value function in stochastic optimal control problems. By the use of the Girsanov change of probability measures, it is demonstrated how a rapidly-exploring random tree (RRT) method can be utilized for the forward integration pass, as long as the controlled drift terms are appropriately compensated in the backward integration pass. Subsequently, a numerical approximation of the value function is proposed by solving a series of function approximation problems backwards in time along the edges of the constructed RRT. Moreover, a local entropy-weighted least squares Monte Carlo (LSMC) method is developed to concentrate function approximation accuracy in regions most likely to be visited by optimally controlled trajectories. The results of the proposed methodology are demonstrated on linear and nonlinear stochastic optimal control problems with non-quadratic running costs, which reveal significant convergence improvements over previous FBSDE-based numerical solution methods.
doi_str_mv 10.48550/arxiv.2006.12444
format Article
fullrecord <record><control><sourceid>arxiv_GOX</sourceid><recordid>TN_cdi_arxiv_primary_2006_12444</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2006_12444</sourcerecordid><originalsourceid>FETCH-LOGICAL-a674-3e35247b23f40e6fbc2408754997c5903bfda12f767c8f73db1cd09525bc59ac3</originalsourceid><addsrcrecordid>eNotj0tOwzAYhL1hgQoHYIUv4OD4ESdLiFpAqqgE2Ue_X2DhxpETQXt70tLVjGak0XwI3ZW0ELWU9AHyIfwUjNKqKJkQ4hq9bVL-hWzJE5jvk8HvMAYbj2R9GGPKYfhcksGmPe6ycxP2KeOPOZkvmOZg8G6cwx4ibtMw5xRv0JWHOLnbi65Qt1l37QvZ7p5f28ctgUoJwh2XTCjNuBfUVV4bJmitpGgaZWRDufYWSuZVpUztFbe6NJY2kkm91GD4Ct3_z56B-jEvH_KxP4H1ZzD-B5hwSNE</addsrcrecordid><sourcetype>Open Access Repository</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Forward-Backward Rapidly-Exploring Random Trees for Stochastic Optimal Control</title><source>arXiv.org</source><creator>Hawkins, Kelsey P ; Pakniyat, Ali ; Theodorou, Evangelos ; Tsiotras, Panagiotis</creator><creatorcontrib>Hawkins, Kelsey P ; Pakniyat, Ali ; Theodorou, Evangelos ; Tsiotras, Panagiotis</creatorcontrib><description>We propose a numerical method for the computation of the forward-backward stochastic differential equations (FBSDE) appearing in the Feynman-Kac representation of the value function in stochastic optimal control problems. By the use of the Girsanov change of probability measures, it is demonstrated how a rapidly-exploring random tree (RRT) method can be utilized for the forward integration pass, as long as the controlled drift terms are appropriately compensated in the backward integration pass. Subsequently, a numerical approximation of the value function is proposed by solving a series of function approximation problems backwards in time along the edges of the constructed RRT. Moreover, a local entropy-weighted least squares Monte Carlo (LSMC) method is developed to concentrate function approximation accuracy in regions most likely to be visited by optimally controlled trajectories. The results of the proposed methodology are demonstrated on linear and nonlinear stochastic optimal control problems with non-quadratic running costs, which reveal significant convergence improvements over previous FBSDE-based numerical solution methods.</description><identifier>DOI: 10.48550/arxiv.2006.12444</identifier><language>eng</language><subject>Computer Science - Robotics ; Computer Science - Systems and Control ; Mathematics - Optimization and Control</subject><creationdate>2020-06</creationdate><rights>http://arxiv.org/licenses/nonexclusive-distrib/1.0</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>228,230,780,885</link.rule.ids><linktorsrc>$$Uhttps://arxiv.org/abs/2006.12444$$EView_record_in_Cornell_University$$FView_record_in_$$GCornell_University$$Hfree_for_read</linktorsrc><backlink>$$Uhttps://doi.org/10.48550/arXiv.2006.12444$$DView paper in arXiv$$Hfree_for_read</backlink></links><search><creatorcontrib>Hawkins, Kelsey P</creatorcontrib><creatorcontrib>Pakniyat, Ali</creatorcontrib><creatorcontrib>Theodorou, Evangelos</creatorcontrib><creatorcontrib>Tsiotras, Panagiotis</creatorcontrib><title>Forward-Backward Rapidly-Exploring Random Trees for Stochastic Optimal Control</title><description>We propose a numerical method for the computation of the forward-backward stochastic differential equations (FBSDE) appearing in the Feynman-Kac representation of the value function in stochastic optimal control problems. By the use of the Girsanov change of probability measures, it is demonstrated how a rapidly-exploring random tree (RRT) method can be utilized for the forward integration pass, as long as the controlled drift terms are appropriately compensated in the backward integration pass. Subsequently, a numerical approximation of the value function is proposed by solving a series of function approximation problems backwards in time along the edges of the constructed RRT. Moreover, a local entropy-weighted least squares Monte Carlo (LSMC) method is developed to concentrate function approximation accuracy in regions most likely to be visited by optimally controlled trajectories. The results of the proposed methodology are demonstrated on linear and nonlinear stochastic optimal control problems with non-quadratic running costs, which reveal significant convergence improvements over previous FBSDE-based numerical solution methods.</description><subject>Computer Science - Robotics</subject><subject>Computer Science - Systems and Control</subject><subject>Mathematics - Optimization and Control</subject><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2020</creationdate><recordtype>article</recordtype><sourceid>GOX</sourceid><recordid>eNotj0tOwzAYhL1hgQoHYIUv4OD4ESdLiFpAqqgE2Ue_X2DhxpETQXt70tLVjGak0XwI3ZW0ELWU9AHyIfwUjNKqKJkQ4hq9bVL-hWzJE5jvk8HvMAYbj2R9GGPKYfhcksGmPe6ycxP2KeOPOZkvmOZg8G6cwx4ibtMw5xRv0JWHOLnbi65Qt1l37QvZ7p5f28ctgUoJwh2XTCjNuBfUVV4bJmitpGgaZWRDufYWSuZVpUztFbe6NJY2kkm91GD4Ct3_z56B-jEvH_KxP4H1ZzD-B5hwSNE</recordid><startdate>20200622</startdate><enddate>20200622</enddate><creator>Hawkins, Kelsey P</creator><creator>Pakniyat, Ali</creator><creator>Theodorou, Evangelos</creator><creator>Tsiotras, Panagiotis</creator><scope>AKY</scope><scope>AKZ</scope><scope>GOX</scope></search><sort><creationdate>20200622</creationdate><title>Forward-Backward Rapidly-Exploring Random Trees for Stochastic Optimal Control</title><author>Hawkins, Kelsey P ; Pakniyat, Ali ; Theodorou, Evangelos ; Tsiotras, Panagiotis</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-a674-3e35247b23f40e6fbc2408754997c5903bfda12f767c8f73db1cd09525bc59ac3</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2020</creationdate><topic>Computer Science - Robotics</topic><topic>Computer Science - Systems and Control</topic><topic>Mathematics - Optimization and Control</topic><toplevel>online_resources</toplevel><creatorcontrib>Hawkins, Kelsey P</creatorcontrib><creatorcontrib>Pakniyat, Ali</creatorcontrib><creatorcontrib>Theodorou, Evangelos</creatorcontrib><creatorcontrib>Tsiotras, Panagiotis</creatorcontrib><collection>arXiv Computer Science</collection><collection>arXiv Mathematics</collection><collection>arXiv.org</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Hawkins, Kelsey P</au><au>Pakniyat, Ali</au><au>Theodorou, Evangelos</au><au>Tsiotras, Panagiotis</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Forward-Backward Rapidly-Exploring Random Trees for Stochastic Optimal Control</atitle><date>2020-06-22</date><risdate>2020</risdate><abstract>We propose a numerical method for the computation of the forward-backward stochastic differential equations (FBSDE) appearing in the Feynman-Kac representation of the value function in stochastic optimal control problems. By the use of the Girsanov change of probability measures, it is demonstrated how a rapidly-exploring random tree (RRT) method can be utilized for the forward integration pass, as long as the controlled drift terms are appropriately compensated in the backward integration pass. Subsequently, a numerical approximation of the value function is proposed by solving a series of function approximation problems backwards in time along the edges of the constructed RRT. Moreover, a local entropy-weighted least squares Monte Carlo (LSMC) method is developed to concentrate function approximation accuracy in regions most likely to be visited by optimally controlled trajectories. The results of the proposed methodology are demonstrated on linear and nonlinear stochastic optimal control problems with non-quadratic running costs, which reveal significant convergence improvements over previous FBSDE-based numerical solution methods.</abstract><doi>10.48550/arxiv.2006.12444</doi><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier DOI: 10.48550/arxiv.2006.12444
ispartof
issn
language eng
recordid cdi_arxiv_primary_2006_12444
source arXiv.org
subjects Computer Science - Robotics
Computer Science - Systems and Control
Mathematics - Optimization and Control
title Forward-Backward Rapidly-Exploring Random Trees for Stochastic Optimal Control
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-02T18%3A29%3A39IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-arxiv_GOX&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Forward-Backward%20Rapidly-Exploring%20Random%20Trees%20for%20Stochastic%20Optimal%20Control&rft.au=Hawkins,%20Kelsey%20P&rft.date=2020-06-22&rft_id=info:doi/10.48550/arxiv.2006.12444&rft_dat=%3Carxiv_GOX%3E2006_12444%3C/arxiv_GOX%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_id=info:pmid/&rfr_iscdi=true