Skip navigation
Skip navigation

Hypothesis testing for management: Evolving and answering closed questions using multiobjective visualization

Kasprzyk, Joseph; Guillaume, Joseph; Kollat, Joshua; Danilo, Chris

Description

In order to use models to understand deeply uncertain future conditions, managers must be able to pose and test hypotheses about their management problems. In Iterative Closed Question Methodology (ICQM), a series of closed questions are used to structure thinking about hypotheses while looking beyond a problem's existing modeling representation. Our research is exploring how ICQM can contribute to a framework called Many Objective Robust Decision Making (MORDM), which uses multiobjective...[Show more]

dc.contributor.authorKasprzyk, Joseph
dc.contributor.authorGuillaume, Joseph
dc.contributor.authorKollat, Joshua
dc.contributor.authorDanilo, Chris
dc.coverage.spatialSan Diego USA
dc.date.accessioned2015-12-10T22:25:16Z
dc.date.createdJune 15-19 2014
dc.identifier.urihttp://hdl.handle.net/1885/53402
dc.description.abstractIn order to use models to understand deeply uncertain future conditions, managers must be able to pose and test hypotheses about their management problems. In Iterative Closed Question Methodology (ICQM), a series of closed questions are used to structure thinking about hypotheses while looking beyond a problem's existing modeling representation. Our research is exploring how ICQM can contribute to a framework called Many Objective Robust Decision Making (MORDM), which uses multiobjective optimization and ensembles of uncertain future states of the world to create and evaluate robust solutions for environmental management. A visualization software tool; AeroVis, has greatly aided implementation of MORDM, allowing a user to plot tradeoffs between conflicting objectives, "brush" their preferences on plotted and unplotted variables, and view visualizations of solution robustness. This visualization approach provides a rich set of conclusions which is not always well understood (i.e. the user can interpret results that the modeler did not intend). In this presentation, we explore how visualization tools iteratively generate and evaluate management hypotheses and conclusions. We discuss the types of conclusions that can be made from AeroVis MORDM visualizations and walk through experimental examples of how individuals reason with the decision support tool. This illustrates that working within an MORDM framework helps the user consider alternate model assumptions about future inputs, parameters and model structure, supporting the idea that model assumptions can provide useful scenarios for environmental management.
dc.publisherConference Organising Committee
dc.relation.ispartofseries7th International Congress on Environmental Modelling and Software, iEMSs 2014
dc.rightsAuthor/s retain copyright
dc.sourceProceedings - 7th International Congress on Environmental Modelling and Software: Bold Visions for Environmental Modeling, iEMSs 2014
dc.titleHypothesis testing for management: Evolving and answering closed questions using multiobjective visualization
dc.typeConference paper
local.description.notesImported from ARIES
local.description.refereedYes
dc.date.issued2014
local.identifier.absfor050205 - Environmental Management
local.identifier.absfor080110 - Simulation and Modelling
local.identifier.ariespublicationa383154xPUB273
local.type.statusPublished Version
local.contributor.affiliationKasprzyk, Joseph, University of Colorado Boulder
local.contributor.affiliationGuillaume, Joseph, College of Medicine, Biology and Environment, ANU
local.contributor.affiliationKollat, Joshua, DecisionVis
local.contributor.affiliationDanilo, Chris, DecisionVis
local.bibliographicCitation.startpage727
local.bibliographicCitation.lastpage733
local.identifier.absseo960913 - Water Allocation and Quantification
dc.date.updated2015-12-09T09:22:29Z
local.identifier.scopusID2-s2.0-84911867801
dcterms.accessRightsOpen Access
CollectionsANU Research Publications

Download

File Description SizeFormat Image
01_Kasprzyk_Hypothesis_testing_for_2014.pdf236.89 kBAdobe PDF


Items in Open Research are protected by copyright, with all rights reserved, unless otherwise indicated.

Updated:  22 January 2019/ Responsible Officer:  University Librarian/ Page Contact:  Library Systems & Web Coordinator