IBM's T. J. Watson Research Center

The following sections are included:Approximating Expected Utility by a Function of Mean and VarianceA Class of ApproximationsAnalysis of Error FunctionsEmpirical ResultsSome Objections ReconsideredThe E,V InvestorREFERENCESMean-variance Versus Direct Utility MaximizationThe ProblemThe Quality of the ApproximationThe Selected Utility FunctionsThe DataThe Empirical ResultsThe Effect of LeverageConclusionsAppendix The Direct Maximization AlgorithmThe ProblemDetails on the Maximization Method of Step 2REFERENCESThe Value of a Blank CheckCONCEPTS OF A BLANK CHECK LOTTERYA SURVEYA PROBLEM WITH EXPONENTIAL UTILITYTHE EFFECTS OF HUMAN CAPITALAPPLICATION TO EVALUATING MEAN-VARIANCE APPROXIMATIONSSUMMARY AND CONCLUSIONAPPENDIXENDNOTESREFERENCESThe “two beta” trapNORMATIVE AND POSITIVE MEAN-VARIANCE ANALYSISPROPOSITIONS ABOUT BETASEPILOGUE AND EDITORIALREFERENCESPortfolio Analysis with Factors and ScenariosThe ModelThe Usual States of the World ModelModeling the ɛ's as being Conditionally UncorrelatedCombining Scenarios and FactorsDiagonalizationConclusionREFERENCESSPARSITY AND PIECEWISE LINEARITY IN LARGE PORTFOLIO OPTIMIZATION PROBLEMSINTRODUCTIONMODELS FOR THE ESTIMATION OF COVARIANCESEXISTING OPTIMIZATION METHODSREVIEW OF THE COMPLEMENTARY PIVOT ALGORITHMGetting startedBasis changesEXPLOITING THE STRUCTURE OF CUse with a sparse matrix packageImplementation out-of-coreIMPLICIT TREATMENT OF UPPER AND LOWER BOUNDS AND TRANSACTIONS COSTSCONCLUSIONS AND SUMMARYREFERENCESDISCUSSIONThe ER and EAS Formalisms for System Modeling and the EAS–E LanguageINTRODUCTIONTHE EAS WORLDVIEWEAS and ERTHE EAS-E LANGUAGETHE REST OF THE ICEBERGTHE EAS PRINCIPLEAPPLICABILITY TO ERREFERENCESEAS-E: An Integrated Approach to Application DevelopmentINTRODUCTIONTHE EAS MODELTHE EAS-E PROGRAMMING LANGUAGESets as Standard Data StructuresIntegrated LanguageSome Examples of EAS-E SyntaxThe FOR EACH PhraseThe FIND StatementFull-Screen Input/Output CapabilityEntity, Attribute, and Set DefinitionsA PL-I/SQL Program and an Equivalent EAS-E ProgramA PLAIN Program and an Equivalent EAS-E ProgramTHE BROWSE FACILITYSOME IMPLEMENTATION DETAILSDatabase Ranked SetsReference and Identifier VariablesRecording and UnlockingTHE EAS-E DATABASE MANAGEMENT SYSTEMConcurrency ControlProtecting Against Software CrashesProtecting Against Physical DamageSUMMARY AND STATUSREFERENCESThe System Architecture of EAS-E: An Integrated Programming and Database LanguageVery little is required to go from the conceptual model to the application programTo write a program that works with the data base, the user must specify which entity types are to be maniplatedSimple queries can be written as small EAS-E programsThe problem of passing the selection information at execution time to the loop-searching mechanism has been addressed by designing an EAS structure to contain that informationEAS-E has been designed to accommodate data bases of arbitrary size, from very small to very largeSamuelson and Investment for the Long RunBackgroundThe Expected Log Rule in General and ParticularFirst Argument For Max E logArgument Against Max E logExampleAnother Argument For Max E logSummaryReferencesINVESTMENT FOR THE LONG RUN: NEW EVIDENCE FOR AN OLD RULEBackgroundThe Sequence of GamesAlternate Sequence-of-Games FormalizationsUnending GamesConclusionsAppendixReferences