Approximate representation of optimal strategies from influence diagrams

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

3 Citations (Scopus)

Abstract

There are three phases in the life of a decision problem, specification, solution, and rep-

resentation of solution. The specification and solution phases are off-line, while the rep-

resention of solution often shall serve an on-line situation with rather tough constraints

on time and space. One of the advantages of influence diagrams (IDs) is that for small

decision problems, the distinction between phases does not confront the decision maker

with a problem; when the problem has been properly specified, the solution algorithms are

so efficient that the ID can also be used as an on-line representation of the solution. If the

solution algorithm cannot meet the on-line requirements, you will construct an alternative

structure for representing the optimal strategy, for example a look-up table or a strategy

tree. We report on ongoing work with situations where the solution algorithm is too space

and time consuming, and where the policy functions for the decisions have so large do-

mains that they cannot be represented directly in a strategy tree. The approach is to have

separate ID representations for each decision variable. In each representation the actual

information is fully exploited, however the representation of policies for future decisions

are approximations. We call the approximation information abstraction. It consists in

introducing a dummy structure connecting the past with the decision. We study how to

specify, implement and learn information abstraction.

Original languageEnglish
Title of host publicationProceedings of the 4th European Workshop on Probabilistic Graphical Models
EditorsManfred Jaeger, Thomas D. Nielsen
Number of pages7
Publication date2008
Pages153-159
Publication statusPublished - 2008
EventPGM08 - Hirtshals, Denmark
Duration: 17 Sep 200819 Sep 2008
Conference number: 4

Conference

ConferencePGM08
Number4
CountryDenmark
CityHirtshals
Period17/09/200819/09/2008

Fingerprint

Specifications

Keywords

  • Influence diagrams

Cite this

Jensen, F. V. (2008). Approximate representation of optimal strategies from influence diagrams. In M. Jaeger, & T. D. Nielsen (Eds.), Proceedings of the 4th European Workshop on Probabilistic Graphical Models (pp. 153-159)
Jensen, Finn V. / Approximate representation of optimal strategies from influence diagrams. Proceedings of the 4th European Workshop on Probabilistic Graphical Models. editor / Manfred Jaeger ; Thomas D. Nielsen. 2008. pp. 153-159
@inproceedings{812e18a0907811dd93c5000ea68e967b,
title = "Approximate representation of optimal strategies from influence diagrams",
abstract = "There are three phases in the life of a decision problem, specification, solution, and rep-resentation of solution. The specification and solution phases are off-line, while the rep-resention of solution often shall serve an on-line situation with rather tough constraintson time and space. One of the advantages of influence diagrams (IDs) is that for smalldecision problems, the distinction between phases does not confront the decision makerwith a problem; when the problem has been properly specified, the solution algorithms areso efficient that the ID can also be used as an on-line representation of the solution. If thesolution algorithm cannot meet the on-line requirements, you will construct an alternativestructure for representing the optimal strategy, for example a look-up table or a strategytree. We report on ongoing work with situations where the solution algorithm is too spaceand time consuming, and where the policy functions for the decisions have so large do-mains that they cannot be represented directly in a strategy tree. The approach is to haveseparate ID representations for each decision variable. In each representation the actualinformation is fully exploited, however the representation of policies for future decisionsare approximations. We call the approximation information abstraction. It consists inintroducing a dummy structure connecting the past with the decision. We study how tospecify, implement and learn information abstraction.",
keywords = "Influence diagrams",
author = "Jensen, {Finn V.}",
year = "2008",
language = "English",
pages = "153--159",
editor = "Manfred Jaeger and Nielsen, {Thomas D.}",
booktitle = "Proceedings of the 4th European Workshop on Probabilistic Graphical Models",

}

Jensen, FV 2008, Approximate representation of optimal strategies from influence diagrams. in M Jaeger & TD Nielsen (eds), Proceedings of the 4th European Workshop on Probabilistic Graphical Models. pp. 153-159, Hirtshals, Denmark, 17/09/2008.

Approximate representation of optimal strategies from influence diagrams. / Jensen, Finn V.

Proceedings of the 4th European Workshop on Probabilistic Graphical Models. ed. / Manfred Jaeger; Thomas D. Nielsen. 2008. p. 153-159.

Research output: Contribution to book/anthology/report/conference proceedingArticle in proceedingResearchpeer-review

TY - GEN

T1 - Approximate representation of optimal strategies from influence diagrams

AU - Jensen, Finn V.

PY - 2008

Y1 - 2008

N2 - There are three phases in the life of a decision problem, specification, solution, and rep-resentation of solution. The specification and solution phases are off-line, while the rep-resention of solution often shall serve an on-line situation with rather tough constraintson time and space. One of the advantages of influence diagrams (IDs) is that for smalldecision problems, the distinction between phases does not confront the decision makerwith a problem; when the problem has been properly specified, the solution algorithms areso efficient that the ID can also be used as an on-line representation of the solution. If thesolution algorithm cannot meet the on-line requirements, you will construct an alternativestructure for representing the optimal strategy, for example a look-up table or a strategytree. We report on ongoing work with situations where the solution algorithm is too spaceand time consuming, and where the policy functions for the decisions have so large do-mains that they cannot be represented directly in a strategy tree. The approach is to haveseparate ID representations for each decision variable. In each representation the actualinformation is fully exploited, however the representation of policies for future decisionsare approximations. We call the approximation information abstraction. It consists inintroducing a dummy structure connecting the past with the decision. We study how tospecify, implement and learn information abstraction.

AB - There are three phases in the life of a decision problem, specification, solution, and rep-resentation of solution. The specification and solution phases are off-line, while the rep-resention of solution often shall serve an on-line situation with rather tough constraintson time and space. One of the advantages of influence diagrams (IDs) is that for smalldecision problems, the distinction between phases does not confront the decision makerwith a problem; when the problem has been properly specified, the solution algorithms areso efficient that the ID can also be used as an on-line representation of the solution. If thesolution algorithm cannot meet the on-line requirements, you will construct an alternativestructure for representing the optimal strategy, for example a look-up table or a strategytree. We report on ongoing work with situations where the solution algorithm is too spaceand time consuming, and where the policy functions for the decisions have so large do-mains that they cannot be represented directly in a strategy tree. The approach is to haveseparate ID representations for each decision variable. In each representation the actualinformation is fully exploited, however the representation of policies for future decisionsare approximations. We call the approximation information abstraction. It consists inintroducing a dummy structure connecting the past with the decision. We study how tospecify, implement and learn information abstraction.

KW - Influence diagrams

M3 - Article in proceeding

SP - 153

EP - 159

BT - Proceedings of the 4th European Workshop on Probabilistic Graphical Models

A2 - Jaeger, Manfred

A2 - Nielsen, Thomas D.

ER -

Jensen FV. Approximate representation of optimal strategies from influence diagrams. In Jaeger M, Nielsen TD, editors, Proceedings of the 4th European Workshop on Probabilistic Graphical Models. 2008. p. 153-159