ebook img

Stochastic Optimization Methods: Applications in Engineering and Operations Research PDF

368 Pages·2015·3.34 MB·English
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview Stochastic Optimization Methods: Applications in Engineering and Operations Research

Kurt Marti Stochastic Optimization Methods Applications in Engineering and Operations Research Third Edition Stochastic Optimization Methods Kurt Marti Stochastic Optimization Methods Applications in Engineering and Operations Research Third edition 123 KurtMarti InstituteforMathematicsandComputer Science FederalArmedForcesUniversityMunich Neubiberg/Munich Germany ISBN978-3-662-46213-3 ISBN978-3-662-46214-0 (eBook) DOI10.1007/978-3-662-46214-0 LibraryofCongressControlNumber:2015933010 SpringerHeidelbergNewYorkDordrechtLondon ©Springer-VerlagBerlinHeidelberg2015 Thisworkissubjecttocopyright.AllrightsarereservedbythePublisher,whetherthewholeorpartof thematerialisconcerned,specificallytherightsoftranslation,reprinting,reuseofillustrations,recitation, broadcasting,reproductiononmicrofilmsorinanyotherphysicalway,andtransmissionorinformation storageandretrieval,electronicadaptation,computersoftware,orbysimilarordissimilarmethodology nowknownorhereafterdeveloped. Theuseofgeneraldescriptivenames,registerednames,trademarks,servicemarks,etc.inthispublication doesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfromtherelevant protectivelawsandregulationsandthereforefreeforgeneraluse. Thepublisher,theauthorsandtheeditorsaresafetoassumethattheadviceandinformationinthisbook arebelievedtobetrueandaccurateatthedateofpublication.Neitherthepublishernortheauthorsor theeditorsgiveawarranty,expressorimplied,withrespecttothematerialcontainedhereinorforany errorsoromissionsthatmayhavebeenmade. Printedonacid-freepaper Springer-VerlagGmbHBerlinHeidelbergispartofSpringerScience+Business Media(www.springer. com) Preface Optimization problems in practice depend mostly on several model parameters, noise factors, uncontrollableparameters,etc., which are not givenfixed quantities attheplanningstage.Typicalexamplesfromengineeringandeconomics/operations research are: material parameters(e.g., elasticity moduli,yield stresses, allowable stresses, moment capacities, specific gravity), external loadings, friction coeffi- cients, moments of inertia, length of links, mass of links, location of center of gravityoflinks,manufacturingerrors,tolerances,noiseterms,demandparameters, technological coefficients in input–output functions, cost factors, interest rates, exchange rates, etc. Due to the several types of stochastic uncertainties (physical uncertainty,economicuncertainty,statisticaluncertainty,modeluncertainty),these parameters must be modeled by random variables having a certain probability distribution.Inmostcases,atleastcertainmomentsofthisdistributionareknown. In order to cope with these uncertainties, a basic procedure in the engi- neering/economic practice is to replace first the unknown parameters by some chosen nominal values, e.g., estimates, guesses, of the parameters. Then, the resulting and mostly increasing deviation of the performance (output, behavior) of the structure/system from the prescribed performance (output, behavior), i.e., the “tracking error”, is compensated by (online) input corrections. However, the onlinecorrectionofasystem/structureisoftentime-consumingandcausesmostly increasing expenses (correction or recourse costs). Very large recourse costs may ariseincaseofdamagesorfailuresoftheplant.Thiscanbeomittedtoalargeextent by taking into account already at the planning stage the possible consequences of the tracking errors and the known prior and sample information about the random data of the problem. Hence, instead of relying on ordinary deterministic parameteroptimizationmethods—basedon some nominalparametervalues—and applyingthenjustsomecorrectionactions,stochasticoptimizationmethodsshould be applied: Incorporating stochastic parameter variations into the optimization process, expensiveand increasingonline correctionexpensescan be omittedor at leastreducedtoalargeextent. Consequently, for the computation of robust optimal decisions/designs, i.e., optimaldecisionswhichareinsensitivewithrespecttorandomparametervariations, v vi Preface appropriatedeterministicsubstituteproblemsmustbeformulatedfirst.Basedonthe decisiontheoreticalprinciples,thesesubstituteproblemsdependonprobabilitiesof failure/successand/oronmoregeneralexpectedcost/lossterms.Sinceprobabilities and expectations are defined by multiple integrals in general, the resulting often nonlinear and also nonconvexdeterministic substitute problems can be solved by approximativemethodsonly.Twobasictypesofdeterministicsubstituteproblems occurmostlyinpractice: • Expected primary cost minimization subject to expected recourse (correction) costconstraints Minimization of the expected primary costs subject to expected recourse costconstraints(reliabilityconstraints)andremainingdeterministicconstraints, e.g., boxconstraints. In case of piecewise constantcost functions,probabilistic objectivefunctionsand/orprobabilisticconstraintsoccur; • ExpectedTotalCostMinimizationProblems: Minimization of the expected total costs (costs of construction, design, recoursecosts,etc.)subjecttotheremainingdeterministicconstraints. Themainanalyticalpropertiesofthesubstituteproblemshavebeenexaminedin thefirsttwoeditionsofthebook,wherealsoappropriatedeterministicandstochastic solutionprocedurescanbefound. Afteranoverviewonbasicmethodsforhandlingoptimizationproblemswithran- domdata,inthispresentthirdeditiontransformationmethodsforoptimizationprob- lems with random parameters into appropriate deterministic substitute problems are described for basic technical and economic problems. Hence, the aim of this present third edition is to provide now analytical and numerical tools—including theirmathematicalfoundations—fortheapproximatecomputationofrobustoptimal decisions/designs/controlasneededinconcreteengineering/economicapplications: stochastic Hamiltonian method for optimal control problems with random data, the H-minimal control and the Hamiltonian two-point boundary value problem, Stochastic optimal open-loop feedback control as an efficient method for the constructionof optimal feedback controlsin case of random parameters, adaptive optimal stochastic trajectory planning and control for dynamic control systems under stochastic uncertainty, optimal design of regulators in case of random parameters, optimal design of structures under random external load and with randommodelparameters. Finally,anewtoolispresentedfortheevaluationoftheentropyofaprobability distribution and the divergence of two probability distributions with respect to the use in an optimal decision problem with random parameters. Applications to statisticsaregiven. Realizing this monograph, the author was supported by several collaborators fromtheInstituteforMathematicsandComputerSciences.Especially,Iwouldlike to thank Ms Elisabeth Lößl for the excellent support in the LATEX-typesetting of the manuscript.Withouther veryprecise and carefulwork, the completionof this projectcouldnothavebeenrealized.Moreover,IthankMsInaSteinforproviding several figures. Last but not least, I am indebted to Springer-Verlag for inclusion Preface vii ofthebookintotheSpringer-program.I wouldliketo thankespeciallythe Senior EditorforBusiness/EconomicsofSpringer-Verlag,ChristianRauscher,forhisvery longpatienceuntilthecompletionofthisbook. Munich,Germany KurtMarti October2014 Preface to the First Edition Optimization problems in practice depend mostly on several model parameters, noise factors, uncontrollableparameters,etc., which are not givenfixed quantities attheplanningstage.Typicalexamplesfromengineeringandeconomics/operations research are: material parameters(e.g., elasticity moduli,yield stresses, allowable stresses, moment capacities, specific gravity), external loadings, friction coeffi- cients, moments of inertia, length of links, mass of links, location of center of gravityoflinks,manufacturingerrors,tolerances,noiseterms,demandparameters, technological coefficients in input–output functions, cost factors, etc. Due to the severaltypesofstochasticuncertainties(physicaluncertainty,economicuncertainty, statistical uncertainty, model uncertainty), these parameters must be modeled by random variables having a certain probability distribution. In most cases, at least certainmomentsofthisdistributionareknown. In order to cope with these uncertainties, a basic procedure in the engi- neering/economic practice is to replace first the unknown parameters by some chosen nominal values, e.g., estimates, guesses, of the parameters. Then, the resulting and mostly increasing deviation of the performance (output, behavior) of the structure/system from the prescribed performance (output, behavior), i.e., the “tracking error,” is compensated by (online) input corrections. However, the onlinecorrectionofasystem/structureisoftentime-consumingandcausesmostly increasing expenses (correction or recourse costs). Very large recourse costs may ariseincaseofdamagesorfailuresoftheplant.Thiscanbeomittedtoalargeextent by taking into account already at the planning stage the possible consequences of the tracking errors and the known prior and sample information about the random data of the problem. Hence, instead of relying on ordinary deterministic parameteroptimizationmethods—basedon some nominalparametervalues—and applyingthenjustsomecorrectionactions,stochasticoptimizationmethodsshould be applied:incorporatingthe stochastic parametervariationsinto the optimization process, expensiveand increasingonline correctionexpensescan be omittedor at leastreducedtoalargeextent. Consequently, for the computation of robust optimal decisions/designs, i.e., optimaldecisionswhichareinsensitivewithrespecttorandomparametervariations, ix

Description:
Features optimization problems that in practice involve random model parameters Provides applications from the fields of robust optimal control / design in case of stochastic uncertainty Includes numerous references to stochastic optimization, stochastic programming and its applications to engineeri
See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.