Resumé

Decision making subject to uncertain information, whether fake or factual, in the context of management of socio-technical systems, is critically discussed from both philosophical and operational perspectives. In dealing with possible fake, incorrect and/or factual information we take the perspective that any information utilized as basis for supporting decisions, has to be dealt with in exactly the same manner - in accordance with Bayesian decision analysis. The important issue is to identify and model the scenarios through which information may cause adverse consequences and to account for their potential effects on the system representation applied as basis for decision optimization. To this end we first provide a mapping of how information affects the decision making context and a categorization of causes for information leading to adverse consequences. Secondly, we introduce a decision analytical framework aiming to optimize decision alternatives for managing systems in the context of systems representations including not only one possible system model but a set of different possible system models. As a means for assessing the benefit of collecting additional information, we utilize Value of Information analysis from Bayesian decision analysis. Finally, a principal example is provided which illustrates and discusses selected aspects of how possibly fake information affects decision making and how it might be dealt with.
OriginalsprogEngelsk
TidsskriftCivil Engineering and Environmental Systems
ISSN1028-6608
DOI
StatusUdgivet - 2019

Fingerprint

Decision theory
Decision making
Information analysis

Citer dette

@article{b00e657ef7e74db5a29b45ee52899342,
title = "Faith and fakes – dealing with critical information in decision analysis",
abstract = "Decision making subject to uncertain information, whether fake or factual, in the context of management of socio-technical systems, is critically discussed from both philosophical and operational perspectives. In dealing with possible fake, incorrect and/or factual information we take the perspective that any information utilized as basis for supporting decisions, has to be dealt with in exactly the same manner - in accordance with Bayesian decision analysis. The important issue is to identify and model the scenarios through which information may cause adverse consequences and to account for their potential effects on the system representation applied as basis for decision optimization. To this end we first provide a mapping of how information affects the decision making context and a categorization of causes for information leading to adverse consequences. Secondly, we introduce a decision analytical framework aiming to optimize decision alternatives for managing systems in the context of systems representations including not only one possible system model but a set of different possible system models. As a means for assessing the benefit of collecting additional information, we utilize Value of Information analysis from Bayesian decision analysis. Finally, a principal example is provided which illustrates and discusses selected aspects of how possibly fake information affects decision making and how it might be dealt with.",
keywords = "risk-informed decision support, Bayesian decision analysis, fake news",
author = "Linda Nielsen and Glavind, {Sebastian T{\o}lb{\o}ll} and Jianjun Qin and Faber, {Michael Havbro}",
year = "2019",
doi = "10.1080/10286608.2019.1615476",
language = "English",
journal = "Civil Engineering and Environmental Systems",
issn = "1028-6608",
publisher = "Taylor & Francis",

}

TY - JOUR

T1 - Faith and fakes – dealing with critical information in decision analysis

AU - Nielsen, Linda

AU - Glavind, Sebastian Tølbøll

AU - Qin, Jianjun

AU - Faber, Michael Havbro

PY - 2019

Y1 - 2019

N2 - Decision making subject to uncertain information, whether fake or factual, in the context of management of socio-technical systems, is critically discussed from both philosophical and operational perspectives. In dealing with possible fake, incorrect and/or factual information we take the perspective that any information utilized as basis for supporting decisions, has to be dealt with in exactly the same manner - in accordance with Bayesian decision analysis. The important issue is to identify and model the scenarios through which information may cause adverse consequences and to account for their potential effects on the system representation applied as basis for decision optimization. To this end we first provide a mapping of how information affects the decision making context and a categorization of causes for information leading to adverse consequences. Secondly, we introduce a decision analytical framework aiming to optimize decision alternatives for managing systems in the context of systems representations including not only one possible system model but a set of different possible system models. As a means for assessing the benefit of collecting additional information, we utilize Value of Information analysis from Bayesian decision analysis. Finally, a principal example is provided which illustrates and discusses selected aspects of how possibly fake information affects decision making and how it might be dealt with.

AB - Decision making subject to uncertain information, whether fake or factual, in the context of management of socio-technical systems, is critically discussed from both philosophical and operational perspectives. In dealing with possible fake, incorrect and/or factual information we take the perspective that any information utilized as basis for supporting decisions, has to be dealt with in exactly the same manner - in accordance with Bayesian decision analysis. The important issue is to identify and model the scenarios through which information may cause adverse consequences and to account for their potential effects on the system representation applied as basis for decision optimization. To this end we first provide a mapping of how information affects the decision making context and a categorization of causes for information leading to adverse consequences. Secondly, we introduce a decision analytical framework aiming to optimize decision alternatives for managing systems in the context of systems representations including not only one possible system model but a set of different possible system models. As a means for assessing the benefit of collecting additional information, we utilize Value of Information analysis from Bayesian decision analysis. Finally, a principal example is provided which illustrates and discusses selected aspects of how possibly fake information affects decision making and how it might be dealt with.

KW - risk-informed decision support, Bayesian decision analysis, fake news

U2 - 10.1080/10286608.2019.1615476

DO - 10.1080/10286608.2019.1615476

M3 - Journal article

JO - Civil Engineering and Environmental Systems

JF - Civil Engineering and Environmental Systems

SN - 1028-6608

ER -