ARGONNENATIONALLABORATORY 9700SouthCassAvenue Argonne,Illinois BenchmarkingOptimizationSoftwarewithPerformanceProles

Size: px
Start display at page:

Download "ARGONNENATIONALLABORATORY 9700SouthCassAvenue Argonne,Illinois60439. BenchmarkingOptimizationSoftwarewithPerformanceProles"

Transcription

1 ARGONNENATIONALLABORATORY 9700SouthCassAvenue Argonne,Illinois60439 BenchmarkingOptimizationSoftwarewithPerformanceProles MathematicsandComputerScienceDivision ElizabethD.DolanandJorgeJ.More PreprintANL/MCS-P January2001 Energy,underContractW Eng-38,andbytheNationalScienceFoundation(ChallengesinComputationalScience)grantCDA and(InformationTechnologyResearch)grantCCR ThisworkwassupportedbytheMathematical,Information,andComputationalSciences DivisionsubprogramoftheOceofAdvancedScienticComputing,U.S.Departmentof

2 Contents 4CaseStudy:OptimalControlandParameterEstimationProblems 2PerformanceEvaluation 3BenchmarkingData 1Introduction CaseStudy:TheFullCOPS 6CaseStudy:LinearProgramming 7Conclusions Acknowledgments References 12

3 BenchmarkingOptimizationSoftwarewithPerformanceProles ElizabethD.DolanyandJorgeJ.Morez 1Introduction Weproposeperformanceproles distributionfunctionsforaperformancemetric asatoolforbenchmarkingandcomparingoptimizationsoftware.weshowthatperformanceprolescombinethebestfeaturesofothertoolsforperformanceevaluation. Abstract withtheevaluationandperformanceofoptimizationcodes.asrecentexamples,wecite [1,2,3,4,6,12,17]. Mittelmann'seortshavegainedthemostnotice,otherresearchershavebeenconcerned Thebenchmarkingofoptimizationsoftwarehasrecentlygainedconsiderablevisibility.Hans themaintechnicalissuesaddressedinthispaper.mostbenchmarkingeortsinvolvetables Mittlemann's[13]workonavarietyofoptimizationsoftwarehasfrequentlyuncovered displayingtheperformanceofeachsolveroneachproblemforasetofmetricssuchascpu decienciesinthesoftwareandhasgenerallyledtosoftwareimprovements.although time,numberoffunctionevaluations,oriterationcountsforalgorithmswhereaniteration impliesacomparableamountofwork.failuretodisplaysuchtablesforasmalltestset wouldbeagrossomission,buttheytendtobeoverwhelmingforlargetestsets.inallcases, theinterpretationoftheresultsfromthesetablesisoftenasourceofdisagreement. Theinterpretationandanalysisofthedatageneratedbythebenchmarkingprocessare researcherstotryvarioustoolsforanalyzingthedata.thesolver'saverageorcumulative mance[1,4,6].asaresult,asmallnumberofthemostdicultproblemscantendto dominatetheseresults,andresearchersmusttakepainstogiveadditionalinformation.anotherdrawbackisthatcomputingaveragesortotalsforaperformancemetricnecessitates discardingproblemsforwhichanysolverfailed,eectivelybiasingtheresultsagainstthe totalforeachperformancemetricoverallproblemsissometimesusedtoevaluateperfor- Thequantitiesofdatathatresultfrombenchmarkingwithlargetestsetshavespurred subprogramoftheoceofadvancedscienticcomputing,u.s.departmentofenergy,undercontract W Eng-38,andbytheNationalScienceFoundation(ChallengesinComputationalScience)grant valuecanbeassignedforfailedsolverattempts,butthisrequiresasubjectivechoicefor mostrobustsolvers.asanalternativetodisregardingsomeoftheproblems,apenalty separatetable. thepenalty.mostresearcherschoosetoreportthenumberoffailuresonlyinafootnoteor CDA and(InformationTechnologyResearch)grantCCR ThisworkwassupportedbytheMathematical,Information,andComputationalSciencesDivision zmathematicsandcomputersciencedivision,argonnenationallaboratory,argonne,illinois60439 ydepartmentofelectricalandcomputerengineering,northwesternuniversity,andmathematicsand 1

4 dierencebetweensolvertimes[4])appearstobeaviablewayofensuringthataminority place,usuallyfork=1;2;3.rankingthesolvers'performanceforeachproblemhelps [4,6,15,17].Inotherwords,theycountthenumberoftimesthatasolvercomesinkth preventaminorityoftheproblemsfromundulyinuencingtheresults.informationonthe sizeoftheimprovement,however,islost. Toaddresstheshortcomingsofthepreviousapproach,someresearchersrankthesolvers oftheproblemsdonotdominatetheresults,butinourtestingwehavewitnessedlargeleaps inquartilevaluesofaperformancemetric,ratherthangradualtrends.ifonlyquartiledata isused,theninformationontrendsoccurringbetweenonequartileandthenextislost;and wemustassumethatthejourneyfromonepointtoanotherproceedsatamoderatepace. Also,inthespeciccaseofcontrastingthedierencesbetweensolvertimes,thecomparison Comparingthemediansandquartilesofsomeperformancemetric(forexample,the isthatifresultsaremixed,interpretingquartiledatamaybenoeasierthanusingtheraw data;anddealingwithcomparisonsofmorethantwosolversmightbecomeunwieldy. failstoprovideanyinformationontherelativesizeoftheimprovement.analdrawback denethebordersofverycompetitiveandcompetitive. appearsin[2],withsolversratedbythepercentageofproblemsforwhichasolver'stimeis termedverycompetitiveorcompetitive.theratioapproachavoidsmostofthediculties thatwehavediscussed,providinginformationonthepercentimprovementandeliminating thenegativeeectsofallowingasmallportionoftheproblemstodominatetheconclusions. Themaindisadvantageofthisapproachliesintheauthor'sarbitrarychoiceoflimitsto Theideaofcomparingsolversbytheratioofonesolver'sruntimetothebestruntime mulative)distributionfunctionforaperformancemetric.inthispaperweusetheratioof thecomputingtimeofthesolverversusthebesttimeofallofthesolversastheperformance metric.section3providesananalysisofthetestsetandsolversusedinthebenchmark theperformanceofoptimizationsoftware.theperformanceproleforasolveristhe(cu- resultsofsections4and5.thisanalysisisnecessarytounderstandthelimitationsofthe benchmarkingprocess. InSection2,weintroduceperformanceprolesasatoolforevaluatingandcomparing parameterchoicesandtheneedtodiscardsolverfailuresfromtheperformancedata. inuenceofasmallnumberofproblemsonthebenchmarkingprocessandthesensitivityof resultsassociatedwiththerankingofsolvers.performanceprolesprovideameansofvisualizingtheexpectedperformancedierenceamongmanysolvers,whileavoidingarbitrary withversion2.0ofthecops[7]testset.weshowthatperformanceproleseliminatethe WeconcludeinSection6byshowinghowperformanceprolesapplytothedata[13] Sections4and5demonstratetheuseofperformanceproleswithresults[8]obtained theuseofperformanceprolesandalsoshowsthatperformanceprolescanbeappliedto ofmittelmannforlinearprogrammingsolvers.thissectionprovidesanothercasestudyof awiderangeofperformancedata. 2

5 informationofinterestsuchasthenumberoffunctionevaluationsandthecomputingtime. Inthissectionweintroducethenotionofaperformanceproleasameanstoevaluateand comparetheperformanceofthesolversonatestsetp. BenchmarkresultsaregeneratedbyrunningasolveronasetPofproblemsandrecording 2PerformanceEvaluation measures.foreachproblempandsolvers,wedene putingtimeasaperformancemeasure;although,theideasbelowcanbeusedwithother Weassumethatwehavenssolversandnpproblems.Weareinterestedinusingcom- performanceratio bysolverswiththebestperformancebyanysolveronthisproblem;thatis,weusethe If,forexample,thenumberoffunctionevaluationsistheperformancemeasureofinterest, settp;saccordingly. Werequireabaselineforcomparisons.Wecomparetheperformanceonproblemp tp;s=computingtimerequiredtosolveproblempbysolvers. solversdoesnotsolveproblemp.theperformanceofsolversonanygivenproblemmay solver.ifwedene beofinterest,butwewouldliketoobtainanoverallassessmentoftheperformanceofthe WeassumethataparameterMp;sforallp;sischosen,andp;s=Mifandonlyif ps()=1npsizenp2p:p;so; p;s= minftp;s:1snsg: possibleratio.thefunctionpsisthe(cumulative)distributionfunctionfortheperformance problemsthatarelikelytooccurinapplications,thensolverswithlargeprobabilityps() ratioẇeusethetermperformanceproleforthedistributionfunctionofaperformancemetric.ourclaimisthataplotoftheperformanceprolerevealsallofthemajorperformance characteristics.inparticular,ifthesetofproblemspissuitablylargeandrepresentativeof thenps()istheprobabilitythataperformanceratiop;siswithinafactorofthebest aretobepreferred. versusaproblemparameter.forexample,higham[11,pages296{297]plotstheratio =kak1,whereistheestimateforthel1normofamatrixaproducedbythelapack conditionnumberestimator.notethatinhigham'suseofthetermperformanceprole thereisnoattemptatdeterminingadistributionfunction. function,continuousfromtherightateachbreakpoint.thevalueofps(1)istheprobability Thetermperformanceprolehasalsobeenusedforaplotofsomeperformancemetric numberofwins,weneedonlytocomparethevaluesofps(1)forallofthesolvers. thatthesolverwillwinovertherestofthesolvers.thus,ifweareinterestedonlyinthe Theperformanceproleps:R7![0;1]forasolverisanondecreasing,piecewiseconstant 3

6 istheprobabilitythatthesolversolvesaproblem.thus,ifweareinterestedonlyinsolvers resultofthisconvention,ps(m)=1,and thatp;s2[1;m]andthatp;s=monlywhenproblempisnotsolvedbysolvers.asa Thedenitionoftheperformanceproleforlargevaluesrequiressomecare.Weassume withahighprobabilityofsuccess,thenweneedtocomparethevaluesofps( M)forall solversandchoosethesolverswiththelargestvalue.thevalueofps( M)canbereadilyseen for2[s;m)forsomes<m. inaperformanceprolebecausepsatlinesforlargevaluesof;thatis,ps()=ps( M) ps( M)lim! Mps() asourceofdisagreementbecausethereisnoconsensusonhowtochooseproblems.thecops problemsareselectedtobeinterestinganddicult,butthiscriteriaissubjective.because 3BenchmarkingData themainpurposeofourtestingistouncoverthebestgeneralsolveronawiderangeof problems,weincludeproblemswithmultipleminimaandproblemswithalargenumberof ThetimingdatausedtocomputetheperformanceprolesinSections4and5isgenerated degreesoffreedom.foreachoftheapplicationsinthecopssetweusefourinstancesofthe withthecopstestset,whichcurrentlyconsistsofseventeendierentapplications,all applicationobtainedbyvaryingaparameterintheapplication,forexample,thenumber modelsintheampl[9]modelinglanguage.thechoiceofthetestproblemsetpisalways mationapplicationsinthecopsset,whilethediscussioninsection5coversthecomplete ofgridpointsinadiscretization. performanceresults.accordingly,weprovidehereinformationspecictothissubsetofthe COPSproblemsaswellasananalysisofthetestsetasawhole.Table3.1givesthequartiles forthreeproblemparameters:thenumberofvariablesn,thenumberofconstraints,and theratio(n ne)=n,whereneisthenumberofequalityconstraints.intheoptimization literature,n neisoftencalledthedegreesoffreedomoftheproblem,sinceitisanupper Section4focusesononlythesubsetoftheelevenoptimalcontrolandparameteresti- throughoutthetestsetandshowsthatatleastthree-fourthsoftheproblemshavethe boundonthenumberofvariablesthatarefreeatthesolution. AnotherfeatureoftheCOPSsubsetisthattheequalityconstraintsaretheresultofeither intherange[1;50]becauseotherbenchmarkingproblemsetstendtohaveapreponderance numberofvariablesnintheinterval[400;5000].ouraimwastoavoidproblemswherenwas ofproblemswithninthisrange.themaindierencebetweenthefullcopssetandthe COPSsubsetisthattheCOPSsubsetismoreconstrainedwithnen=2foralltheproblems. ThedatainTable3.1isfairlyrepresentativeofthedistributionoftheseparameters settingtheoutputlevelsothatwecangatherthedataweneed,increasingtheiteration dierenceorcollocationapproximationstodierentialequations. Weranournalcompleterunswiththesameoptionsforallmodels.Theoptionsinvolve 4

7 to5000.noneofthefailureswerecordinthenaltrialsincludeanysolvererrormessages abouthavingviolatedtheselimits. limitsasmuchasallowed,andincreasingthesuper-basicslimitsforminosandsnopt Numberofconstraints Degreesoffreedom(%) Numberofvariables minq1q2q3maxminq1q2q3max Table3.1:ProblemdataforCOPStestset FullCOPS COPSsubset soastominimizetheeectofuctuationinthemachineload.thescripttracksthewallclocktimefromthestartofthesolve,killinganyprocessthatruns3,600seconds,which recordingthewall-clocktimeaswellasthecombinationofamplsystemtime(tointerpret themodelandcomputevaryingamountsofderivativeinformationrequiredbyeachsolver) andsolvertimeforeachmodelvariation.werepeatthecycleforanymodelforwhichone wedeclareunsuccessful,andbeginningthenextsolve.wecyclethroughalltheproblems, Thescriptforgeneratingthetimingdatasendsaproblemtoeachsolversuccessively, ofthesolver'sampltimesandthewall-clocktimesdierbymorethan10percent.to furtherensureconsistency,wehaveveriedthattheampltimeresultswepresentcouldbe requirementscanalsoplayanimportantrole.inparticular,solversthatusedirectlinear criteriaarenotlikelytochangecomputingtimesbymorethanafactoroftwo.memory intheamplenvironment.inanycase,dierencesincomputingtimeduetothestopping thesolvers.ideallywewouldhaveusedthesamestoppingcriteria,butthisisnotpossible runningsolaris7. reproducedtowithin10percentaccuracy.allcomputationsweredoneonasparcultra2 equationsolversareoftenmoreecientintermsofcomputingtimeprovidedthereisenough memory. Wehaveignoredtheeectsofthestoppingcriteriaandthememoryrequirementsof mation.theuseofsecond-orderinformationcanreducethenumberofiterations,but useonlyrst-orderinformation,whilelancelotandloqoneedsecond-orderinfor- thecostperiterationusuallyincreases.inaddition,obtainingsecond-orderinformationis morecostlyandmaynotevenbepossible.minosandsnoptarespecicallydesigned LANCELOTandLOQO.Asanexampleofcomparingsolverswithsimilarrequirements, forproblemswithamodestnumberofdegreesoffreedom,whilethisisnotthecasefor Thesolversthatwebenchmarkherehavedierentrequirements.MINOSandSNOPT Section6showstheperformanceoflinearprogrammingsolvers. 5

8 4CaseStudy:OptimalControlandParameterEstimationProblems WenowexaminetheperformanceprolesofLANCELOT[5],MINOS[14],SNOPT[10], andloqo[16]onthesubsetoftheoptimalcontrolandparameterestimationproblemsin thecops[7]testset.figures4.1and4.2showtheperformanceprolesindierentranges providethenecessaryobjectiveinformationforreasonablesubjectiveanalysisofalargetest set.figure4.1showstheperformanceprolesofthefoursolversforsmallvaluesof. Byshowingtheratiosofsolvertimes,weeliminateanyweightofimportancethattaking toreectvariousareasofinterest.ourpurposeistoshowhowtheperformanceproles ofbeingtheoptimalsolver)andthattheprobabilitythatloqoisthewinneronagiven straighttimedierencesmightgivetotheproblemsthatrequirealongruntimeofevery solversreceivetheirduecreditforcompletingproblemsforwhichoneormoreoftheother solversfails.inparticular,1 ps()isthefractionofproblemsthatthesolvercannotsolve withinafactorofthebestsolver,includingproblemsforwhichthesolverinquestionfails. solver.wendnoneedtoeliminateanytestproblemsfromdiscussion.forthisreason, showsthattheprobabilitythatthesetwosolverscansolveajobwithinafactor4ofthe MINOS,butitsperformancebecomesmuchmorecompetitiveifweextendourofinterest bestsolverisonlyabout70%.snopthasalowernumberofwinsthaneitherloqoor problemisabout:61.ifwechoosebeingwithinafactorof4ofthebestsolverasthescope ofourinterest,theneitherloqoorminoswouldsuce;buttheperformanceprole FromthisgureitisclearthatLOQOhasthemostwins(hasthehighestprobability successfully,thensnoptcapturesourattentionwithitsabilitytosolveover90%ofthis areinterestedinthesolverthatcansolve75%oftheproblemswiththegreatesteciency, thenminosstandsout.ifweholdtomorestringentprobabilitiesofcompletingasolve COPSsubset,asdisplayedbytheheightofitsperformanceprolefor>40.Thisgraph to7ḟigure4.2showstheperformanceprolesforallthesolversintheinterval[1;100].ifwe interval[15;40]. percentageoftheproblems.anotherpointofinterestisthatloqo,minos,andsnopt eachhavethebestprobabilityps()forinsomeinterval,withsimilarperformanceinthe displaysthepotentialforlargediscrepanciesintheperformanceratiosonasubstantial betweenquartilesdoesnotnecessarilyproceedlinearly;hence,wereallyloseinformation forsnoptand13:9forloqo.bylookingatfigures4.1and4.2,weseethatprogress MINOSalsosharerstquartilevaluesof1.Inotherwords,thesetwosolversarethe bestsolversonatleast25%oftheproblems.loqobestsminos'smedianvaluewith1 comparedwith2:4,butminoscomesbackwithathirdquartileratioof4:3versus12:6 valuesoftimeratios.thetopthreesolversshareaminimumratioof1,andloqoand Anobservationthatemergesfromtheseguresisthelackofconsistencyinquartile andnoobviousalternativevalueexists.asanalternativetoprovidingonlyquartilevalues, ifwedonotprovidethefulldata.also,themaximumratiowouldbemforourtesting, 6

9 1 Performance Profile on Subset of COPS P( ( t p,s / min {t p,s : 1 s n s } ) τ) 0.7 Figure4.1:Performanceproleon[0;10] LANCELOT MINOS 0.3 SNOPT LOQO τ 1 Performance Profile on Subset of COPS 0.9 P( ( t p,s / min {t p,s : 1 s n s } ) τ) LANCELOT MINOS SNOPT τ Figure4.2:Performanceproleon[0;100]LOQO

10 however,theperformanceproleyieldsmuchmoreinformationaboutasolver'sstrengths andweaknesses. of2forthescale.inotherwords,weplot fullimplicationsofourtestdataregardingthesolvers'probabilityofsuccessfullyhandling aproblem.sincewearealsointerestedinthebehaviorforclosetounity,weuseabase thesolvers.evenextendingto100,wefailtocapturethecompleteperformancedata proles.inthisway,wecanshowallactivitythattakesplacewith<mandgraspthe forlancelotandloqo.asanaloption,wedisplayalogscaleoftheperformance Wehaveseenthatatleasttwographsmaybeneededtoexaminetheperformanceof thegraphisnotasintuitive,sinceweareusingalogscale. infigure4.3.thisgraphrevealsallthefeaturesoftheprevioustwographsandthus capturestheperformanceofallthesolvers.thedisadvantageisthattheinterpretationof 7!1npsizefp2P:log2(p;s)g 1 Log 2 Scaled Performance Profile on Subset of COPS 0.9 P( log 2 ( t p,s / min {t p,s : 1 s n s } ) τ) interval[0;log2(1043)]infigure4.3toincludethelargestp;s<m.weextendtherange theothertwoguresand,inaddition,showsthateachofthesolversfailsonatleast8%of slightlytoshowtheatliningofallsolvers.thenewgurecontainsalltheinformationof Figures4.1and4.2aremappedintoanewscaletoreectalldata,requiringatleastthe Figure4.3:Performanceproleinalog2scale LANCELOT 0.2 MINOS 8 SNOPT LOQO ρ M τ

11 problemsweregenerallychosentobedicult. theproblems.thisisnotanunreasonableperformanceforthecopstestsetbecausethese 5CaseStudy:TheFullCOPS 1 Log 2 Scaled Performance Profile on Full COPS Set 0.9 P( log 2 ( t i,j / min {t i,j : 1 j n s } ) τ) theperformanceprolesinsection4,thisgureshowsthatperformanceproleseliminate Figure5.1:PerformanceproleforfullCOPSset 0.2 theundueinuenceofasmallnumberofproblemsonthebenchmarkingprocessandthe LANCELOT sensitivityoftheresultsassociatedwiththerankingofsolvers.inaddition,performance Wenowexpandouranalysisofthedatatoincludealltheproblemsinversion2.0ofthe MINOS 0.1 SNOPT prolesprovideanestimateoftheexpectedperformancedierencebetweensolvers. COPS[7]testset.WepresentinFigure5.1alog2scaledviewoftheperformanceproles forthesolversonthattestset. Figure5.1givesaclearindicationoftherelativeperformanceofeachsolver.Asin LOQO ρ testsetloqodominatesallothersolvers:theperformanceproleforloqoliesabove ThemostsignicantaspectofFigure5.1,ascomparedwithFigure4.3,isthatonthis M τ allothersforallperformanceratios.theinterpretationoftheresultsinfigure5.1is important.inparticular,theseresultsdonotimplythatloqoisfasteroneveryproblem. Theyindicateonlythat,forany1,LOQOsolvesmoreproblemswithinafactorof ofanyothersolvertime.moreover,byexaminingps(1)andps(m),wecanalsosaythat 9

12 LOQOisthefastestsolveronapproximately58%oftheproblems,andthatLOQOsolves themostproblems(about87%)tooptimality. Figure5.1withFigure4.3andnotingthattheperformanceprolesofMINOSandSNOPT SNOPTtodeteriorateonthefullCOPSset.Thisdeteriorationcanbeseenbycomparing setismuchlargerthanfortherestrictedsubsetofoptimalcontrolandparameterestimation problems.since,asnotedinsection3,minosandsnoptaredesignedforproblemswith amodestnumberofdegreesoffreedom,weshouldexpecttheperformanceofminosand factors.firstofall,ascanbeseenintable3.1,thedegreesoffreedomforthefullcopstest ThedierencebetweentheresultsinSection4andtheseresultsisduetoanumberof thatminosandsnoptuseonlyrst-orderinformation,whileloqousessecond-order information.thebenetofusingsecond-orderinformationusuallyincreasesasthenumber ofvariablesincreases,sothisisanotherfactorthatbenetsloqo. aconvenienttoolforcomparingandevaluatingtheperformanceofoptimizationsolvers,but, aresimilarbutgenerallylowerinfigure5.1. likealltools,performanceprolesmustbeusedwithcare.aperformanceprolereects Theresultsinthissectionunderscoreourobservationthatperformanceprolesprovide AnotherreasonforthedierencebetweentheresultsinSection4andtheseresultsis theperformanceonlyonthedatabeingused,andthusitisimportanttounderstandthe testsetandthesolversusedinthebenchmark. casestudy,weusedataobtainedbymittelmann[13].figure6.1showsaplotoftheperformanceproleforthetimeratiosinthedatabenchmarkoflpsolversonalinux-pmanceprolesaremostusefulincomparingseveralsolvers.becauselargeamountsofdata aregeneratedinthesesituations,trendsinperformanceareoftendiculttosee.asa 6CaseStudy:LinearProgramming ( ),whichincludesresultsforCOPLLP-1.0,PCx-1.1,SOPLEX-1.1,LPABO-5.6, Performanceprolescanbeusedtocomparetheperformanceoftwosolvers,butperfor- withoutconvergenceunderthesolver'sstoppingcriteria.onefeatureweseeinthegraphof Mittelmann'sresultsthatdoesnotappearintheCOPSgraphsisthevisualdisplayofsolvers thatneveratline.inotherwords,thesolversthatclimbothegrapharethosethatsolve allofthetestproblemssuccessfully.aswithfigure4.3,alloftheeventsinthedatatinto thosesolvesthataremarkedintheoriginaltableasstoppingclosetothenalsolution MOSEK-1.0b,BPMPD-2.11,andBPMPD thislog-scaledrepresentation.whilethisdatasetcannotbeuniversallyrepresentativeof InkeepingwithourgraphingpracticeswiththeCOPSset,wedesignateasfailures particular,thetestsetusedtogeneratefigure6.1includesonlyproblemsselectedby benchmarkingresultsbyanymeans,itdoesshowthatourreportingtechniqueisapplicable beyondourownresults. oftheperformanceoflpsolversonlyonthedatasetusedtogeneratetheseresults.in AsinthecasestudiesinSections4and5,theresultsinFigure6.1giveanindication 10

13 1 Log 2 Scaled Performance Profile for LP Solvers 0.9 BPMPD 2.14 P( log 2 ( t p,s / min {t p,s : 1 s n s } ) τ) 0.8 MOSEK 0.7 PCx 0.6 SOPLEXC LPABO 0.5 BPMPD 2.11 Mittelmannforhisbenchmark.Theadvantageoftheseresultsisthat,unlikethesolversin COPL_LP 0.4 Sections4and5,allsolversinFigure6.1havethesamerequirements Conclusions Figure6.1:Performanceproleforlinearprogrammingsolvers 0.2 Wehaveshownthatperformanceprolescombinethebestfeaturesofothertoolsforbenchmarkingandcomparingoptimizationsolvers.Clearly,theuseofperformanceprolesisnot 0.1 restrictedtooptimizationsolversandcanbeusedtocomparesolversinotherareas. 0 ThePerlscriptperf.plontheCOPSsite[7]generatesperformanceprolesformatted ρ M τ adjustedwithinmatlabtoreectparticularbenchmarkinginterests. toshowthefullareaofactivity.theareadisplayedandscaleofthegraphcanthenbe asmatlabcommandstoproduceacompositegraphasinfigures4.1and4.2.thescript acceptsdataforseveralsolversandplotstheperformanceproleonanintervalcalculated 11

14 Acknowledgments Version1.0oftheCOPSproblemswasdevelopedbyAlexanderBondarenko,DavidBortz, LizDolan,andMichaelMerritt.Theircontributionswereessentialbecause,inmanycases, References thewayforcopswithhisbenchmarkingwork. ouslysharedtheiramplexpertisewithus.finally,wethankhansmittelmannforpaving version2.0oftheproblemsarecloselyrelatedtotheoriginalversion. andspiriteddiscussionsonproblemformulation,whilebobfoureranddavidgaygener- AlexBondarenko,NickGould,SvenLeyerandBobVanderbeicontributedinteresting [1]H.Y.Benson,D.F.Shanno,andR.J.Vanderbei,Interior-pointmethodsfor [3]A.S.Bondarenko,D.M.Bortz,andJ.J.More,COPS:Large-scalenonlinearlyconstrainedoptimizationproblems,TechnicalMemorandumANL/MCS-TM-237, ArgonneNationalLaboratory,Argonne,Illinois,1998(RevisedOctober1999). large-scalemixedcomplementarityproblems,comp.optim.appl.,7(1997),pp.3{25. nicalreportorfe-00-02,princetonuniversity,princeton,newjersey,2000. [2]S.C.Billups,S.P.Dirkse,andM.C.Ferris,Acomparisonofalgorithmsfor nonconvexnonlinearprogramming:jammingandcomparativenumericaltesting,tech- [5]A.R.Conn,N.I.M.Gould,andP.L.Toint,LANCELOT,no.17inSpringer [6],NumericalexperimentswiththeLANCELOTpackage(ReleaseA)forlarge-scale [4]I.Bongartz,A.R.Conn,N.I.M.Gould,M.A.Saunders,andP.L.Toint, AnumericalcomparisonbetweentheLANCELOTandMINOSpackagesforlarge-scale numericaloptimization,report97/13,namuruniversity,1997. [8]E.D.DolanandJ.J.More,BenchmarkingoptimizationsoftwarewithCOPS, [7]COPS.Seehttp:// SeriesinComputationalMathematics,Springer-Verlag,1992. TechnicalMemorandumANL/MCS-TM-246,ArgonneNationalLaboratory,Argonne, Illinois,2000. nonlinearoptimization,math.programming,73(1996),pp.73{110. [10]P.E.Gill,W.Murray,andM.A.Saunders,SNOPT:Analgorithmforlargescaleconstrainedoptimization,ReportNA97-2,UniversityofCalifornia,SanDiego, MathematicalProgramming,TheScienticPress,1993. [9]R.Fourer,D.M.Gay,andB.W.Kernighan,AMPL:AModelingLanguagefor

15 [11]N.J.Higham,AccuracyandStabilityofNumericalAlgorithms,SIAM,Philadelphia, [12]H.Mittelmann,BenchmarkinginteriorpointLP/QPsolvers,Opt.Meth.Software, Pennsylvania,1996. [14]B.A.MurtaghandM.A.Saunders,MINOS5.4user'sguide,ReportSOL83-20R, [13],Benchmarksforoptimizationsoftware,2000.Seehttp://plato.la.asu.edu/ bench.html. StanfordUniversity, (1999),pp.655{670. [17]R.J.VanderbeiandD.F.Shanno,Aninterior-pointalgorithmfornonconvex [16]R.J.Vanderbei,LOQOuser'smanual{Version3.10,Opt.Meth.Software,12 [15]S.G.NashandJ.Nocedal,AnumericalstudyofthelimitedmemoryBFGSmethod (1999),pp.485{514. andthetruncatednewtonmethodforlargescaleoptimization,siamj.optim.,1 (1991),pp.358{372. nonlinearprogramming,comp.optim.appl.,13(1999),pp.231{

An Overview Of Software For Convex Optimization. Brian Borchers Department of Mathematics New Mexico Tech Socorro, NM 87801 borchers@nmt.

An Overview Of Software For Convex Optimization. Brian Borchers Department of Mathematics New Mexico Tech Socorro, NM 87801 borchers@nmt. An Overview Of Software For Convex Optimization Brian Borchers Department of Mathematics New Mexico Tech Socorro, NM 87801 [email protected] In fact, the great watershed in optimization isn t between linearity

More information

Conic optimization: examples and software

Conic optimization: examples and software Conic optimization: examples and software Etienne de Klerk Tilburg University, The Netherlands Etienne de Klerk (Tilburg University) Conic optimization: examples and software 1 / 16 Outline Conic optimization

More information

Solving NP Hard problems in practice lessons from Computer Vision and Computational Biology

Solving NP Hard problems in practice lessons from Computer Vision and Computational Biology Solving NP Hard problems in practice lessons from Computer Vision and Computational Biology Yair Weiss School of Computer Science and Engineering The Hebrew University of Jerusalem www.cs.huji.ac.il/ yweiss

More information

Modern Optimization Methods for Big Data Problems MATH11146 The University of Edinburgh

Modern Optimization Methods for Big Data Problems MATH11146 The University of Edinburgh Modern Optimization Methods for Big Data Problems MATH11146 The University of Edinburgh Peter Richtárik Week 3 Randomized Coordinate Descent With Arbitrary Sampling January 27, 2016 1 / 30 The Problem

More information

Finding Liveness Errors with ACO

Finding Liveness Errors with ACO Hong Kong, June 1-6, 2008 1 / 24 Finding Liveness Errors with ACO Francisco Chicano and Enrique Alba Motivation Motivation Nowadays software is very complex An error in a software system can imply the

More information

Uncertainty modeling revisited: What if you don t know the probability distribution?

Uncertainty modeling revisited: What if you don t know the probability distribution? : What if you don t know the probability distribution? Hans Schjær-Jacobsen Technical University of Denmark 15 Lautrupvang, 2750 Ballerup, Denmark [email protected] Uncertain input variables Uncertain system

More information

Discrete Optimization

Discrete Optimization Discrete Optimization [Chen, Batson, Dang: Applied integer Programming] Chapter 3 and 4.1-4.3 by Johan Högdahl and Victoria Svedberg Seminar 2, 2015-03-31 Todays presentation Chapter 3 Transforms using

More information

Dantzig-Wolfe bound and Dantzig-Wolfe cookbook

Dantzig-Wolfe bound and Dantzig-Wolfe cookbook Dantzig-Wolfe bound and Dantzig-Wolfe cookbook [email protected] DTU-Management Technical University of Denmark 1 Outline LP strength of the Dantzig-Wolfe The exercise from last week... The Dantzig-Wolfe

More information

Optimal Hiring of Cloud Servers A. Stephen McGough, Isi Mitrani. EPEW 2014, Florence

Optimal Hiring of Cloud Servers A. Stephen McGough, Isi Mitrani. EPEW 2014, Florence Optimal Hiring of Cloud Servers A. Stephen McGough, Isi Mitrani EPEW 2014, Florence Scenario How many cloud instances should be hired? Requests Host hiring servers The number of active servers is controlled

More information

Performance of Optimization Software - an Update

Performance of Optimization Software - an Update Performance of Optimization Software - an Update INFORMS Annual 2011 Charlotte, NC 13-18 November 2011 H. D. Mittelmann School of Math and Stat Sciences Arizona State University 1 Services we provide Guide

More information

A CP Scheduler for High-Performance Computers

A CP Scheduler for High-Performance Computers A CP Scheduler for High-Performance Computers Thomas Bridi, Michele Lombardi, Andrea Bartolini, Luca Benini, and Michela Milano {thomas.bridi,michele.lombardi2,a.bartolini,luca.benini,michela.milano}@

More information

Bandwidth Allocation in a Network Virtualization Environment

Bandwidth Allocation in a Network Virtualization Environment Bandwidth Allocation in a Network Virtualization Environment Juan Felipe Botero [email protected] Xavier Hesselbach [email protected] Department of Telematics Technical University of Catalonia

More information

GAMS, Condor and the Grid: Solving Hard Optimization Models in Parallel. Michael C. Ferris University of Wisconsin

GAMS, Condor and the Grid: Solving Hard Optimization Models in Parallel. Michael C. Ferris University of Wisconsin GAMS, Condor and the Grid: Solving Hard Optimization Models in Parallel Michael C. Ferris University of Wisconsin Parallel Optimization Aid search for global solutions (typically in non-convex or discrete)

More information

Illustration (and the use of HLM)

Illustration (and the use of HLM) Illustration (and the use of HLM) Chapter 4 1 Measurement Incorporated HLM Workshop The Illustration Data Now we cover the example. In doing so we does the use of the software HLM. In addition, we will

More information

Portfolio Distribution Modelling and Computation. Harry Zheng Department of Mathematics Imperial College [email protected]

Portfolio Distribution Modelling and Computation. Harry Zheng Department of Mathematics Imperial College h.zheng@imperial.ac.uk Portfolio Distribution Modelling and Computation Harry Zheng Department of Mathematics Imperial College [email protected] Workshop on Fast Financial Algorithms Tanaka Business School Imperial College

More information

Gate Delay Model. Estimating Delays. Effort Delay. Gate Delay. Computing Logical Effort. Logical Effort

Gate Delay Model. Estimating Delays. Effort Delay. Gate Delay. Computing Logical Effort. Logical Effort Estimating Delays Would be nice to have a back of the envelope method for sizing gates for speed Logical Effort Book by Sutherland, Sproull, Harris Chapter 1 is on our web page Also Chapter 4 in our textbook

More information

1 3 5 7 9 11 12 13 15 [Nm] 400 375 350 325 300 275 250 225 200 175 150 155 PS 100 PS 125 PS [kw][ps] 140 190 130 176 120 163 110 149 100 136 125 30 100 20 1000 1500 2000 2500 3000 3500 4000 4500 RPM 90

More information

TOMLAB - For fast and robust largescale optimization in MATLAB

TOMLAB - For fast and robust largescale optimization in MATLAB The TOMLAB Optimization Environment is a powerful optimization and modeling package for solving applied optimization problems in MATLAB. TOMLAB provides a wide range of features, tools and services for

More information

Lecture 3. Linear Programming. 3B1B Optimization Michaelmas 2015 A. Zisserman. Extreme solutions. Simplex method. Interior point method

Lecture 3. Linear Programming. 3B1B Optimization Michaelmas 2015 A. Zisserman. Extreme solutions. Simplex method. Interior point method Lecture 3 3B1B Optimization Michaelmas 2015 A. Zisserman Linear Programming Extreme solutions Simplex method Interior point method Integer programming and relaxation The Optimization Tree Linear Programming

More information

Warshall s Algorithm: Transitive Closure

Warshall s Algorithm: Transitive Closure CS 0 Theory of Algorithms / CS 68 Algorithms in Bioinformaticsi Dynamic Programming Part II. Warshall s Algorithm: Transitive Closure Computes the transitive closure of a relation (Alternatively: all paths

More information

Lecture 11: 0-1 Quadratic Program and Lower Bounds

Lecture 11: 0-1 Quadratic Program and Lower Bounds Lecture : - Quadratic Program and Lower Bounds (3 units) Outline Problem formulations Reformulation: Linearization & continuous relaxation Branch & Bound Method framework Simple bounds, LP bound and semidefinite

More information

Minimizing the Number of Machines in a Unit-Time Scheduling Problem

Minimizing the Number of Machines in a Unit-Time Scheduling Problem Minimizing the Number of Machines in a Unit-Time Scheduling Problem Svetlana A. Kravchenko 1 United Institute of Informatics Problems, Surganova St. 6, 220012 Minsk, Belarus [email protected] Frank

More information

Algorithm Design and Analysis

Algorithm Design and Analysis Algorithm Design and Analysis LECTURE 27 Approximation Algorithms Load Balancing Weighted Vertex Cover Reminder: Fill out SRTEs online Don t forget to click submit Sofya Raskhodnikova 12/6/2011 S. Raskhodnikova;

More information

Numerical methods for American options

Numerical methods for American options Lecture 9 Numerical methods for American options Lecture Notes by Andrzej Palczewski Computational Finance p. 1 American options The holder of an American option has the right to exercise it at any moment

More information

Error Log Processing for Accurate Failure Prediction. Humboldt-Universität zu Berlin

Error Log Processing for Accurate Failure Prediction. Humboldt-Universität zu Berlin Error Log Processing for Accurate Failure Prediction Felix Salfner ICSI Berkeley Steffen Tschirpke Humboldt-Universität zu Berlin Introduction Context of work: Error-based online failure prediction: error

More information

Discuss the size of the instance for the minimum spanning tree problem.

Discuss the size of the instance for the minimum spanning tree problem. 3.1 Algorithm complexity The algorithms A, B are given. The former has complexity O(n 2 ), the latter O(2 n ), where n is the size of the instance. Let n A 0 be the size of the largest instance that can

More information

A numerically adaptive implementation of the simplex method

A numerically adaptive implementation of the simplex method A numerically adaptive implementation of the simplex method József Smidla, Péter Tar, István Maros Department of Computer Science and Systems Technology University of Pannonia 17th of December 2014. 1

More information

Airlift: Video Conferencing as a Cloud Service using Inter- Datacenter Networks

Airlift: Video Conferencing as a Cloud Service using Inter- Datacenter Networks Airlift: Video Conferencing as a Cloud Service using Inter- Datacenter Networks Yuan Feng Baochun Li Bo Li University of Toronto HKUST 1 Multi-party video conferencing 2 Multi-party video conferencing

More information

A Tool to aid nmon analysis and utility

A Tool to aid nmon analysis and utility A Tool to aid nmon analysis and utility By David Tansley and Dave Seol Many AIX system administrators and engineers consider nmon as the greatest performance tool created for AIX. They love nmom for being

More information

Matpower 4.1 User s Manual

Matpower 4.1 User s Manual Matpower 4.1 User s Manual Ray D. Zimmerman Carlos E. Murillo-Sánchez December 14, 2011 2010, 2011 Power Systems Engineering Research Center (Pserc) All Rights Reserved Contents 1 Introduction 7 1.1 Background................................

More information

Data Mining on Streams

Data Mining on Streams Data Mining on Streams Using Decision Trees CS 536: Machine Learning Instructor: Michael Littman TA: Yihua Wu Outline Introduction to data streams Overview of traditional DT learning ALG DT learning ALGs

More information

Hardware/Software Codesign

Hardware/Software Codesign Hardware/Software Codesign. Review. Allocation, Binding and Scheduling Marco Platzner Lothar Thiele by the authors Synthesis Behavior Structure Synthesis Tasks Œ Allocation: Œ Binding: Œ Scheduling: selection

More information

Optimal Income Taxation with Multidimensional Types

Optimal Income Taxation with Multidimensional Types 1 Optimal Income Taxation with Multidimensional Types Kenneth L. Judd Hoover Institution NBER Che-Lin Su Northwestern University (Stanford Ph. D.) PRELIMINARY AND INCOMPLETE November, 2006 2 Introduction

More information

Toward Variability Management to Tailor High Dimensional Index Implementations

Toward Variability Management to Tailor High Dimensional Index Implementations Toward Variability Management to Tailor High Dimensional Index Implementations Presented by: Azeem Lodhi 14.05.2015 Veit Köppen & Thorsten Winsemann & 1 Gunter Saake Agenda Motivation Business Data Warehouse

More information

FUZZY CLUSTERING ANALYSIS OF DATA MINING: APPLICATION TO AN ACCIDENT MINING SYSTEM

FUZZY CLUSTERING ANALYSIS OF DATA MINING: APPLICATION TO AN ACCIDENT MINING SYSTEM International Journal of Innovative Computing, Information and Control ICIC International c 0 ISSN 34-48 Volume 8, Number 8, August 0 pp. 4 FUZZY CLUSTERING ANALYSIS OF DATA MINING: APPLICATION TO AN ACCIDENT

More information

Spare part inventory control for an aircraft component repair shop

Spare part inventory control for an aircraft component repair shop Spare part inventory control for an aircraft component repair shop Martin de Jong - Fokker Services Willem van Jaarsveld - Erasmus Universiteit Rotterdam 19th of January 2011 Outline Introduction Component

More information

HYBRID GENETIC ALGORITHMS FOR SCHEDULING ADVERTISEMENTS ON A WEB PAGE

HYBRID GENETIC ALGORITHMS FOR SCHEDULING ADVERTISEMENTS ON A WEB PAGE HYBRID GENETIC ALGORITHMS FOR SCHEDULING ADVERTISEMENTS ON A WEB PAGE Subodha Kumar University of Washington [email protected] Varghese S. Jacob University of Texas at Dallas [email protected]

More information

INTERIOR-POINT METHODS FOR NONCONVEX NONLINEAR PROGRAMMING: FILTER METHODS AND MERIT FUNCTIONS

INTERIOR-POINT METHODS FOR NONCONVEX NONLINEAR PROGRAMMING: FILTER METHODS AND MERIT FUNCTIONS INTERIOR-POINT METHODS FOR NONCONVEX NONLINEAR PROGRAMMING: FILTER METHODS AND MERIT FUNCTIONS HANDE Y. BENSON, DAVID F. SHANNO, AND ROBERT J. VANDERBEI Operations Research and Financial Engineering Princeton

More information

DATA VERIFICATION IN ETL PROCESSES

DATA VERIFICATION IN ETL PROCESSES KNOWLEDGE ENGINEERING: PRINCIPLES AND TECHNIQUES Proceedings of the International Conference on Knowledge Engineering, Principles and Techniques, KEPT2007 Cluj-Napoca (Romania), June 6 8, 2007, pp. 282

More information

Extension of measure

Extension of measure 1 Extension of measure Sayan Mukherjee Dynkin s π λ theorem We will soon need to define probability measures on infinite and possible uncountable sets, like the power set of the naturals. This is hard.

More information

! Solve problem to optimality. ! Solve problem in poly-time. ! Solve arbitrary instances of the problem. !-approximation algorithm.

! Solve problem to optimality. ! Solve problem in poly-time. ! Solve arbitrary instances of the problem. !-approximation algorithm. Approximation Algorithms Chapter Approximation Algorithms Q Suppose I need to solve an NP-hard problem What should I do? A Theory says you're unlikely to find a poly-time algorithm Must sacrifice one of

More information

BUSINESS ANALYTICS. Data Pre-processing. Lecture 3. Information Systems and Machine Learning Lab. University of Hildesheim.

BUSINESS ANALYTICS. Data Pre-processing. Lecture 3. Information Systems and Machine Learning Lab. University of Hildesheim. Tomáš Horváth BUSINESS ANALYTICS Lecture 3 Data Pre-processing Information Systems and Machine Learning Lab University of Hildesheim Germany Overview The aim of this lecture is to describe some data pre-processing

More information

Linear Programming Sensitivity Analysis

Linear Programming Sensitivity Analysis Linear Programming Sensitivity Analysis Massachusetts Institute of Technology LP Sensitivity Analysis Slide 1 of 22 Sensitivity Analysis Rationale Shadow Prices Definition Use Sign Range of Validity Opportunity

More information

Parallelization Strategies for Multicore Data Analysis

Parallelization Strategies for Multicore Data Analysis Parallelization Strategies for Multicore Data Analysis Wei-Chen Chen 1 Russell Zaretzki 2 1 University of Tennessee, Dept of EEB 2 University of Tennessee, Dept. Statistics, Operations, and Management

More information

THE PROBLEM WORMS (1) WORMS (2) THE PROBLEM OF WORM PROPAGATION/PREVENTION THE MINIMUM VERTEX COVER PROBLEM

THE PROBLEM WORMS (1) WORMS (2) THE PROBLEM OF WORM PROPAGATION/PREVENTION THE MINIMUM VERTEX COVER PROBLEM 1 THE PROBLEM OF WORM PROPAGATION/PREVENTION I.E. THE MINIMUM VERTEX COVER PROBLEM Prof. Tiziana Calamoneri Network Algorithms A.y. 2014/15 2 THE PROBLEM WORMS (1)! A computer worm is a standalone malware

More information

Copula model estimation and test of inventory portfolio pledge rate

Copula model estimation and test of inventory portfolio pledge rate International Journal of Business and Economics Research 2014; 3(4): 150-154 Published online August 10, 2014 (http://www.sciencepublishinggroup.com/j/ijber) doi: 10.11648/j.ijber.20140304.12 ISS: 2328-7543

More information

Dušan Bernát ([email protected])

Dušan Bernát (bernat@fiit.stuba.sk) Domain Name System as a Memory and Communication Medium Dušan Bernát ([email protected]) Institute of Computer Systems and Networks, Faculty of Informatics and Information Technology, STU Bratislava,

More information

Lab 5 Linear Regression with Within-subject Correlation. Goals: Data: Use the pig data which is in wide format:

Lab 5 Linear Regression with Within-subject Correlation. Goals: Data: Use the pig data which is in wide format: Lab 5 Linear Regression with Within-subject Correlation Goals: Data: Fit linear regression models that account for within-subject correlation using Stata. Compare weighted least square, GEE, and random

More information

11. APPROXIMATION ALGORITHMS

11. APPROXIMATION ALGORITHMS 11. APPROXIMATION ALGORITHMS load balancing center selection pricing method: vertex cover LP rounding: vertex cover generalized load balancing knapsack problem Lecture slides by Kevin Wayne Copyright 2005

More information

Preprint 2009-02. Ayalew Getachew Mersha, Stephan Dempe Feasible Direction Method for Bilevel Programming Problem ISSN 1433-9307

Preprint 2009-02. Ayalew Getachew Mersha, Stephan Dempe Feasible Direction Method for Bilevel Programming Problem ISSN 1433-9307 Fakultät für Mathematik und Informatik Preprint 2009-02 Ayalew Getachew Mersha, Stephan Dempe Feasible Direction Method for Bilevel Programming Problem ISSN 1433-9307 Ayalew Getachew Mersha, Stephan Dempe

More information

An Additive Neumann-Neumann Method for Mortar Finite Element for 4th Order Problems

An Additive Neumann-Neumann Method for Mortar Finite Element for 4th Order Problems An Additive eumann-eumann Method for Mortar Finite Element for 4th Order Problems Leszek Marcinkowski Department of Mathematics, University of Warsaw, Banacha 2, 02-097 Warszawa, Poland, [email protected]

More information

Mixed Integer Linear Programming in R

Mixed Integer Linear Programming in R Mixed Integer Linear Programming in R Stefan Theussl Department of Statistics and Mathematics Wirtschaftsuniversität Wien July 1, 2008 Outline Introduction Linear Programming Quadratic Programming Mixed

More information

Lecture 5: Logical Effort

Lecture 5: Logical Effort Introduction to CMOS VLSI Design Lecture 5: Logical Effort David Harris Harvey Mudd College Spring 2004 Outline Introduction Delay in a Logic Gate Multistage Logic Networks Choosing the Best Number of

More information

Optimization of IP Load-Balanced Routing for Hose Model

Optimization of IP Load-Balanced Routing for Hose Model 2009 21st IEEE International Conference on Tools with Artificial Intelligence Optimization of IP Load-Balanced Routing for Hose Model Invited Paper Eiji Oki Ayako Iwaki The University of Electro-Communications,

More information

SBB: A New Solver for Mixed Integer Nonlinear Programming

SBB: A New Solver for Mixed Integer Nonlinear Programming SBB: A New Solver for Mixed Integer Nonlinear Programming Michael R. Bussieck GAMS Development Corp. Arne S. Drud ARKI Consulting & Development A/S OR2001, Duisburg Overview! SBB = Simple Branch & Bound!

More information

Probabilistic Models for Big Data. Alex Davies and Roger Frigola University of Cambridge 13th February 2014

Probabilistic Models for Big Data. Alex Davies and Roger Frigola University of Cambridge 13th February 2014 Probabilistic Models for Big Data Alex Davies and Roger Frigola University of Cambridge 13th February 2014 The State of Big Data Why probabilistic models for Big Data? 1. If you don t have to worry about

More information

Load Balancing by MPLS in Differentiated Services Networks

Load Balancing by MPLS in Differentiated Services Networks Load Balancing by MPLS in Differentiated Services Networks Riikka Susitaival, Jorma Virtamo, and Samuli Aalto Networking Laboratory, Helsinki University of Technology P.O.Box 3000, FIN-02015 HUT, Finland

More information

IEOR 4404 Homework #2 Intro OR: Deterministic Models February 14, 2011 Prof. Jay Sethuraman Page 1 of 5. Homework #2

IEOR 4404 Homework #2 Intro OR: Deterministic Models February 14, 2011 Prof. Jay Sethuraman Page 1 of 5. Homework #2 IEOR 4404 Homework # Intro OR: Deterministic Models February 14, 011 Prof. Jay Sethuraman Page 1 of 5 Homework #.1 (a) What is the optimal solution of this problem? Let us consider that x 1, x and x 3

More information

A Service Design Problem for a Railway Network

A Service Design Problem for a Railway Network A Service Design Problem for a Railway Network Alberto Caprara Enrico Malaguti Paolo Toth Dipartimento di Elettronica, Informatica e Sistemistica, University of Bologna Viale Risorgimento, 2-40136 - Bologna

More information

ANALYSIS, THEORY AND DESIGN OF LOGISTIC REGRESSION CLASSIFIERS USED FOR VERY LARGE SCALE DATA MINING

ANALYSIS, THEORY AND DESIGN OF LOGISTIC REGRESSION CLASSIFIERS USED FOR VERY LARGE SCALE DATA MINING ANALYSIS, THEORY AND DESIGN OF LOGISTIC REGRESSION CLASSIFIERS USED FOR VERY LARGE SCALE DATA MINING BY OMID ROUHANI-KALLEH THESIS Submitted as partial fulfillment of the requirements for the degree of

More information

Understanding the Impact of Weights Constraints in Portfolio Theory

Understanding the Impact of Weights Constraints in Portfolio Theory Understanding the Impact of Weights Constraints in Portfolio Theory Thierry Roncalli Research & Development Lyxor Asset Management, Paris [email protected] January 2010 Abstract In this article,

More information

GAMS Productivity - Performance - Reliability

GAMS Productivity - Performance - Reliability GAMS Productivity - Performance - Reliability Jan-H. Jagla, Lutz Westermann GAMS Software GmbH Annual Review Meeting CAPD, CMU Pittsburgh, PA, March 12 13, 2007 Agenda GAMS Productivity Performance Reliability

More information

PuLP: A Linear Programming Toolkit for Python

PuLP: A Linear Programming Toolkit for Python PuLP: A Linear Programming Toolkit for Python Stuart Mitchell, Stuart Mitchell Consulting, Michael O Sullivan, Iain Dunning Department of Engineering Science, The University of Auckland, Auckland, New

More information

Medication Prescribing Practices of CRNPs in Mental Health Programs and Mental Retardation Programs BY:

Medication Prescribing Practices of CRNPs in Mental Health Programs and Mental Retardation Programs BY: OFFICE OF MENTAL HEALTH AND SUBSTANCE ABUSE SERVICES and OFFICE OF DEVELOPMENTAL PROGRAMS ISSUE DATE: March 27, 2010 EFFECTIVE DATE: March 27, 2010 NUMBER: OMHSAS -10-01 SUBJECT: Medication Prescribing

More information

BEFORE THE INSURANCE COMMISSIONER OF THE COMMONWEALTH OF PENNSYLVANIA DECISION AND ORDER

BEFORE THE INSURANCE COMMISSIONER OF THE COMMONWEALTH OF PENNSYLVANIA DECISION AND ORDER BEFORE THE INSURANCE COMMISSIONER OF THE COMMONWEALTH OF PENNSYLVANIA In Re: Application of AF&L Holdings, LLC in Support of the Request for Approval to Acquire Control of AF&L Insurance Company and Senior

More information

5 INTEGER LINEAR PROGRAMMING (ILP) E. Amaldi Fondamenti di R.O. Politecnico di Milano 1

5 INTEGER LINEAR PROGRAMMING (ILP) E. Amaldi Fondamenti di R.O. Politecnico di Milano 1 5 INTEGER LINEAR PROGRAMMING (ILP) E. Amaldi Fondamenti di R.O. Politecnico di Milano 1 General Integer Linear Program: (ILP) min c T x Ax b x 0 integer Assumption: A, b integer The integrality condition

More information

Enterprise Applications in the Cloud: Non-virtualized Deployment

Enterprise Applications in the Cloud: Non-virtualized Deployment Enterprise Applications in the Cloud: Non-virtualized Deployment Leonid Grinshpan, Oracle Corporation (www.oracle.com) Subject The cloud is a platform devised to support a number of concurrently working

More information

Windows Scheduled Task and PowerShell Scheduled Job Management Pack Guide for Operations Manager 2012

Windows Scheduled Task and PowerShell Scheduled Job Management Pack Guide for Operations Manager 2012 Windows Scheduled Task and PowerShell Scheduled Job Management Pack Guide for Operations Manager 2012 Published: July 2014 Version 1.2.0.500 Copyright 2007 2014 Raphael Burri, All rights reserved Terms

More information

Lemma 5.2. Let S be a set. (1) Let f and g be two permutations of S. Then the composition of f and g is a permutation of S.

Lemma 5.2. Let S be a set. (1) Let f and g be two permutations of S. Then the composition of f and g is a permutation of S. Definition 51 Let S be a set bijection f : S S 5 Permutation groups A permutation of S is simply a Lemma 52 Let S be a set (1) Let f and g be two permutations of S Then the composition of f and g is a

More information

EMAIL MARKETING TIP. Ask to share with a friend

EMAIL MARKETING TIP. Ask to share with a friend Better Business Better Life 1. EMAIL MARKETING TIP Ask to share with a friend Concept contributed to this post by Adam Franklin Time To Action: Option 1 1min to action, Option 2 1min to action, Option

More information

HURDLE AND SELECTION MODELS Jeff Wooldridge Michigan State University BGSE/IZA Course in Microeconometrics July 2009

HURDLE AND SELECTION MODELS Jeff Wooldridge Michigan State University BGSE/IZA Course in Microeconometrics July 2009 HURDLE AND SELECTION MODELS Jeff Wooldridge Michigan State University BGSE/IZA Course in Microeconometrics July 2009 1. Introduction 2. A General Formulation 3. Truncated Normal Hurdle Model 4. Lognormal

More information

Fundamentals and Recent Developments of the GAMS System

Fundamentals and Recent Developments of the GAMS System Fundamentals and Recent Developments of the GAMS System Toni Lastusilta [email protected] Alexander Meeraus [email protected] Franz Nelissen [email protected] GAMS Development Corp. / GAMS Software

More information

SOLVING EQUATIONS WITH EXCEL

SOLVING EQUATIONS WITH EXCEL SOLVING EQUATIONS WITH EXCEL Excel and Lotus software are equipped with functions that allow the user to identify the root of an equation. By root, we mean the values of x such that a given equation cancels

More information

Airport Planning and Design. Excel Solver

Airport Planning and Design. Excel Solver Airport Planning and Design Excel Solver Dr. Antonio A. Trani Professor of Civil and Environmental Engineering Virginia Polytechnic Institute and State University Blacksburg, Virginia Spring 2012 1 of

More information

Modeling and Performance Evaluation of Computer Systems Security Operation 1

Modeling and Performance Evaluation of Computer Systems Security Operation 1 Modeling and Performance Evaluation of Computer Systems Security Operation 1 D. Guster 2 St.Cloud State University 3 N.K. Krivulin 4 St.Petersburg State University 5 Abstract A model of computer system

More information