Solving acomprehensivemodelformultiobjectiveprojectportfolioselection

Any organizationisroutinelyfacedwiththeneedtomakedecisionsregardingtheselectionandscheduling of projectportfoliosfromasetofcandidateprojects.Weproposeamultiobjectivebinaryprogramming model thatfacilitatesbothobtainingefficientportfoliosinlinewiththesetofobjectivespursuedbythe organization, aswellastheirschedulingregardingtheoptimumtimetolauncheachprojectwithinthe portfolio withouttheneedforaprioriinformationonthedecision-makerʹspreferences.Resourcecon- straints, thepossibilityoftransferringresourcesnotconsumedinagivenaperiodtothefollowingone, and projectinterdependencehavealsobeentakenintoaccount.Giventhatthecomplexityofthisproblem increases asthenumberofprojectsandthenumberofobjectivesincrease,wesolveitusingametaheuris- tic procedurebasedonScatterSearchthatwecallSS-PPS(ScatterSearchforProjectPortfolioSelection). The characteristicsandeffectivenessofthismethodarecomparedwithotherheuristicapproaches(SPEA and afullyrandomprocedure)usingcomputationalexperimentsonrandomlygeneratedinstances. Statement ofscopeandpurpose This paperdescribesamodeltoaidintheselectionandschedulingof project portfolioswithinanorganization.Themodelwasdesignedassumingstronginterdependence between projects,whichthereforehavetobeassessedingroups,whileallowingindividualprojectsto start atdifferenttimesdependingonresourceavailabilityoranyotherstrategicorpoliticalrequirements, which involvestimingissues.Thesimultaneouscombinationofprojectportfolioselectionandscheduling under generalconditionsinvolvesknowndrawbacksthatweattempttoremedy.Finally,themodeltakes into accountmultipleobjectiveswithoutrequiringapriorispecificationsregardingthedecision-makerʹs preferences.