diplomsko delo univerzitetnega študijskega programa
Marjeta Uršej (Author), Zorka Novak-Pintarič (Mentor), Zdravko Kravanja (Co-mentor)

Abstract

V diplomski nalogi so predstavljeni različni pristopi za obvladovanje tveganja in negotovosti pri sprejemanju odločitev z matematičnim programiranjem. Prikazane metode vključujejo odločitvena drevesa, deterministični pristop, pristop »počakaj in poglej« z domnevo popolne informacije, dvostopenjsko stohastično metodo z rekurzom in simulacijo Monte Carlo. Prikazana sta izračun in pomen pričakovane vrednosti popolne informacije in vrednost stohastične rešitve. Uporaba zgoraj navedenih pristopov je prikazana na petih primerih, ki predstavljajo odločitvene probleme na področjih planiranja proizvodnje, trženja in investiranja v negotovih pogojih. Matematični modeli študijskih primerov spadajo v skupino linearnega programiranja (LP) in mešano celoštevilskega linearnega programiranja (MILP oz. MIP). V primeru večperiodnega planiranja proizvodnje kemijskih obratov je izveden test fleksibilnosti. V primeru večperiodnega planiranja investiranja je izvedeno dvokriterijsko optimiranje z vključitvijo tveganja, izraženega z varianco, v namensko funkcijo stohastičnega modela. S spreminjanjem uteži variance v namenski funkciji je dobljena Pareto krivulja, ki kaže, da se z nižanjem stopnje tveganja, ki jo je odločevalec pripravljen sprejeti, znižuje tudi pričakovana ekonomska korist. Na osnovi rešenih primerov zaključujemo, da metoda, ki temelji na domnevi popolne informacije, ni ustrezno orodje za sprejemanje odločitev. Primernejši pristop je dvostopenjsko stohastično programiranje z rekurzom, ki daje optimalne vrednosti prvostopenjskih spremenljivk, preden je negotovost razrešena, medtem ko se vrednosti drugostopenjskih spremenljivk prilagajajo glede na realizacijo negotovih parametrov v prihodnosti. Na ta način dobimo optimalno kompromisno rešitev, ki je najbolje zavarovana proti tveganju možnih scenarijev.

Keywords

tveganje;negotovost;tehniška ekonomika;stohastično programiranje;večperiodni model;kemijski procesi;investicijsko odločanje;

Data

Language: Slovenian
Year of publishing:
Source: Maribor
Typology: 2.11 - Undergraduate Thesis
Organization: UM FKKT - Faculty of Chemistry and Chemical Engineering
Publisher: [M. Uršej]
UDC: 519.863:66(043.2)
COBISS: 13926678 Link will open in a new window
Views: 2620
Downloads: 179
Average score: 0 (0 votes)
Metadata: JSON JSON-RDF JSON-LD TURTLE N-TRIPLES XML RDFA MICRODATA DC-XML DC-RDF RDF

Other data

Secondary language: English
Secondary title: Uncertainty and risk in engineering economy
Secondary abstract: Diploma paper presents various approaches for managing risk and uncertainty in decision-making process by means of mathematical programming. The approaches include decision trees, deterministic approach, wait & see method with assumption of perfect information, two-stage stochastic problems with recourse, and Monte Carlo simulation. The expected value of perfect information and the value of stochastic solution are presented as well. Five examples are used to illustrate the applications of the above mentioned approaches. These examples describe decision-making problems in production, management, and investment under uncertainty. Mathematical models correspond to linear programming (LP) and mixed integer linear programming (MILP or MIP) models. Flexibility test was carried out in the case of multiperiod planning of chemical complex. In the case of multiperiod investment planning, two-criteria optimization was performed by including risk, expressed as a variance in the objective function of the stochastic program. By varying the weight coefficient of the variance in the objective function, trade-off was established between expected net present value and risk. The generated Pareto curves show that the expected economic benefit decreases with decreasing level of risk acceptable for a decision maker. It could be concluded that wait & see method is not an effective tool for decision making as it relies on the assumption of perfect information. Two-stage stochastic programming with recourse is more appropriate, because it determines optimal values of first stage variables before the uncertainty is revealed. Later after resolving uncertainty, the second-stage variables are determined based on the actual outcome of uncertainty. In this way optimal solution is generated which is hedged against risk of all possible scenarios.
Secondary keywords: risk;engineering economy;stochastic programming;multiperiod model;chemical process;investment decision-making;
URN: URN:SI:UM:
Type (COBISS): Undergraduate thesis
Thesis comment: Univ. v Mariboru, Fak. za kemijo in kemijsko tehnologijo
Pages: XV, 120 str.
Keywords (UDC): mathematics;natural sciences;naravoslovne vede;matematika;mathematics;matematika;operational research (or): mathematical theories and methods;operacijsko raziskovanje;applied sciences;medicine;technology;uporabne znanosti;medicina;tehnika;chemical technology;chemical and related industries;kemijska tehnologija;kemijske in sorodne industrije;
ID: 987775