Suboptimal dynamic programming with error bounds

Summary, in English

This paper presents a method to relax Dynamic Programming. The methodmakes it possible to findsuboptimal solutions with known error bounds to hard problems.The bounds are chosen by the user, who can then effectively trade-offbetween solution time and accuracy. Several examples from differentdomains where the method is highly useful are presented.

• Författare: Bo Lincoln; Anders Rantzer
• Avdelning/ar: Institutionen för reglerteknik
• Publiceringsår: 2002
• Språk: Engelska
• Sidor: 2354-2359
• Publikation/Tidskrift/Serie: Proceedings of the 41st IEEE Conference on Decision and Control, 2002
• Volym: 2
• Fulltext: - 382 kB; Download statistics
• Länkar: http://ieeexplore.ieee.org/iel5/8437/26567/01184885.pdf; http://dx.doi.org/10.1109/CDC.2002.1184885
• Dokumenttyp: Konferensbidrag
• Förlag: IEEE - Institute of Electrical and Electronics Engineers Inc.
• Ämne: Control Engineering
• Nyckelord: optimal control; dynamic programming; suboptimal; pomdp; piecewise linear
• Conference name: 41st IEEE Conference on Decision and Control
• Conference date: 2002-12-10 - 2002-12-13
• Conference place: Las Vegas, NV, United States
• Status: Published
• ISBN/ISSN/Övrigt: ISSN: 0191-2216; ISBN: 0-7803-7516-5