Title Optimizacija raspodjele kontejnera na slagalištu lučkoga kontejnerskog terminala Title (english) Optimisation of container relocation problem in port container terminal Author Livia Maglić Mentor Čedomir Dundović (mentor)Committee member Svjetlana Hess (predsjednik povjerenstva)Committee member Čedomir Dundović (član povjerenstva)Committee member Zvonko Kavran (član povjerenstva)Granter University of Rijeka Defense date and country 2016-02-05, Croatia Scientific / art field, TECHNICAL SCIENCES Universal decimal classificationUDC ) 62 - Engineering. Technology in general 656.6 - Transport by water Abstract U radu je istražen operativni logistički problem razmještaja kontejnera (BRP/CRP) na
slagalištu lučkoga kontejnerskog terminala. Cilj istraživanja bio je razviti optimizacijski
model koji rješava problem razmještaja kontejnera te omogućava brže i lakše donošenje
odluka na operativnoj razini planiranja na kontejnerskom terminalu. Predloženi
optimizacijski model izrađen je primjenom genetskog algoritma koji na osnovi 4
definirana pravila premještaja određuje poziciju premještaja kontejnera unutar odjeljka.
Rezultat optimizacije za promatrani odjeljak predstavlja set pravila premještaja kojim se
postiže minimalan broj radnji premještaja kontejnera unutar odjeljka. Model je testiran
na osnovi 4 veličine odjeljka i 3 različita scenarija u pogledu udjela popunjenosti
odjeljaka. Za svaku veličinu i scenarij popunjenosti odjeljka slučajno je stvoren jedan
početni raspored kontejnera unutar odjeljka. S ciljem utvrđivanja različitosti rješenja s
aspekta dobivenog seta pravila (kromosoma rješenja), odjeljak zadane veličine i
scenarija popunjenosti te istog početnog rasporeda testiran je prilikom deset pokretanja
modela. Dakle, model je ukupno pokrenut 120 puta, a kao zaključak se nameće to da je
za isti početni raspored odjeljka zadane veličine i scenarija popunjenosti odjeljka
moguće dobiti jednako kvalitetno ili optimalno rješenje primjenom različitog seta
pravila tj; premještanjem kontejnera na različite kontejnerske pozicije.
Dodatno testiranje predloženog modela provedeno je s ciljem usporedbe rezultata
modela s rezultatima drugih autora koji su za izradu modela koristili metaheurističke i
heurističke metode. Testiranje predloženog modela provedeno je za 4 veličine odjeljka i
jedan scenarij s udjelom popunjenosti od približno 90% na uzorku od 40 slučajno
stvorenih različitih početnih rasporeda odjeljaka, dakle ukupno je ostvareno 160
pokretanja modela. Dobiveni rezultati pokazuju učinkovitost predloženog modela s
maksimalnim postignutim poboljšanjem od 16,59 %.
Izrađena je i analiza osjetljivosti predloženog modela na promjenu parametara
genetskog algoritma, na osnovi koje je zaključeno da su svi parametri rada genetskog
algoritma postavljeni tako da omogućavaju postizanje optimalnih ili najkvalitetnijih
rješenja.
Abstract (english) This paper analyses the operational logistical problem in the container storage yard,
called the block relocation problem or container relocation problem (BRP/CRP). The
aim of the paper was to develop an optimization model that solves BRP/CRP and
enables faster and easier decision-making at the operational planning level at the
container terminal. The proposed optimization model is based on genetic algorithm
which determines the storage location of moving container based upon 4 relocation
rules. The result of optimization, for a given yard bay, is based upon a set of relocation
rules which minimize the number of relocation movements. The model was tested for a
4 different sizes and 3 different scenarios of a bay occupancy. For each size and yard bay
occupancy one initial layout of containers in the yard bay was generated randomly. In
order to determine the diversity of obtained results in terms of chromosome solution,
given yard bay defined size, occupancy and the same initial layout was tested ten times.
Thus, the model was running total of 120 times and we can conclude that for a given
yard bay defined with size, occupancy and initial layout, it can obtain high-quality
solution or optimal solution by applying different sets of rules. Additional testing of the
proposed model was conducted to compare the results of the model with results of the
other authors who solved CRP/BRP problem by a metaheuristics and heuristics
methodology. Testing of the proposed model was conducted for 4 different yard bay
sizes and one scenario of yard bay occupancy with share of about 90% on a sample of 40
randomly generated different initial yard bay layout. Thus model was running for total
of 160 times. The results shows the effectiveness of the proposed model with the
maximum achieved improvement of 16.59%. A sensitivity analysis of proposed model is
made on which it was concluded that all the parameters of genetic algorithm are set to
ensure high-quality or optimal solution.
Keywords
lučki kontejnerski terminal
slagalište
logistika
problem razmještaja kontejnera
optimizacija
genetski algoritam
Keywords (english)
port container terminal
storage yard
logistics
container relocation problem
optimisation
genetic algorithm
Language croatian URN:NBN urn:nbn:hr:187:051373 Promotion 2020 Study programme Title: Postgraduate (doctoral) university study programme - Maritime Studies Catalog URL http://libraries.uniri.hr/cgi-bin/lb07/unilib.cgi?form=D1100317074 Type of resource Text Extent V, 157 listova : ilustr., graf. prikazi (djelomice u bojama) ; 30 cm File origin Born digital Access conditions Open access Terms of use Created on 2016-10-26 09:33:45
