magistrsko delo
Karmen Balantič (Author), Dejan Dragan (Mentor)

Abstract

Za učinkovito planiranje in sprejemanje pravih odločitev morajo danes podjetja znati predvideti stanje v prihodnosti, zato so metode za napovedovanje nepogrešljive, hkrati pa hitro se spreminjajoče politično in gospodarsko okolje vpliva na to, da osnovne metode za napovedovanje niso več dovolj. Zato smo v magistrskem delu poskušali preučiti genetske algoritme in njihovo uporabnost pri napovedovanju pretovora v Luki Koper, d. d. Izdelali smo dva avtoregresijska integrirana modela drsečih sredin s pojasnjevalnimi spremenljivkami (modela ARIMAX) za napovedovanje pretovora kontejnerjev in napovedovanje razsutega tovora. Pojasnjevalne spremenljivke so nam predstavljali različni makroekonomski kazalniki (bruto domači proizvod, uvoz/izvoz, stopnja brezposelnosti ter pariteta kupnih moči), ki vplivajo na pretovor v Luki Koper. Uporabnost genetskih algoritmov smo v modelu preizkusili dvakrat, prvič za izbiro primernih makroekonomskih kazalnikov kot vhodov ARIMAX modela, kjer smo genetske algoritme združili z regresijo delnih najmanjših kavdratov, ter drugič za izbiro najprimernejšega ARIMAX modela. Dobljena modela sta ustrezala vsem pogojem za stabilnost in ustreznost modela ter dokaj dobro zajela dinamiko časovnih vrst, zaradi česar lahko primernost uporabne genetskih algoritmom pri napovedovanju pretovora potrdimo.

Keywords

genetski algoritmi;modeliranje;model ARIMAX;napovedovanje pretovora;logistika;magistrske naloge;

Data

Language: Slovenian
Year of publishing:
Typology: 2.09 - Master's Thesis
Organization: UM FL - Faculty of Logistics
Publisher: [K. Balantič]
UDC: 004.42
COBISS: 512842557 Link will open in a new window
Views: 963
Downloads: 146
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: Applying genetic algorithms for throughput forecasting at port of Koper
Secondary abstract: Regarding effective planning and decision making companies today must be able to foresee the situation in the future, therefore methods for forecasting are necessary. At the same time quickly changing political and economic environment influences on such manner, that basic methods for forecasting are no longer enough. In this thesis, we wanted to study the genetic algorithms and their use in throughput forecasting at Port of Koper. We have developed two autoregressive integrated moving average models with explanatory variables (ARIMAX model) for forecasting the throughput of containers and bulk. The explanatory variables we used were various macroeconomic indicators (gross domestic product, import/export, the unemployment rate and purchasing power parity), which have the highest impact on throughput at the Port of Koper. We tested the usefulness of genetic algorithms twice. First, we use them for the selection of the best macroeconomic indicators as inputs of ARIMAX model and secondly, for selection of the best ARIMAX model. Obtained results showed that both models fit all the conditions for stability and adequacy and they both fairly well captured the dynamics of time series. On that matter we were able to confirm that the use of genetic algorithms in throughput forecasting is suitable.
Secondary keywords: genetic algorithms;modelling;ARIMAX model;throughput forecasting;logistics;
URN: URN:SI:UM:
Type (COBISS): Master's thesis/paper
Thesis comment: Univ. v Mariboru, Fak. za logistiko
Pages: X, 114 str., [2] str. pril.
ID: 9604837