Simple proceduresofchoiceinmulticriteriaproblemswithoutprecise informationaboutthealternatives’values

The additivemodelofmultiattributevaluetheoryiswidelyusedinmulticriteriachoiceproblems.But often itisnoteasytoobtainprecisevaluesforthescalingweightsorthealternatives’valueineach function.Severaldecisionruleswhichrequireweakerinformation,suchasordinalinformation,have been proposedtoselectanalternativeunderthesecircumstances.Weproposenewdecisionrulesand test themusingMonte-Carlosimulation,consideringthatthereisordinalinformationbothonthe scaling weightsandonthealternatives’values.Resultsshowthenewrulesconstituteagood approximation.Weprovideguidelinesabouthowtousetheserulesinacontextofselectingasubsetof the mostpromisingalternatives,consideringthecontradictoryobjectivesofkeepingalownumberof alternativesyetnotexcludingthebestone.