321
1 INTRODUCTION
The analytical techniques for the assessment of
elements’ parameters to big systems with complex
functionalconnections[1‐2].Thedemandsforgetting
more adequate and accurate estimations of
characteristics,translatedintoselectedcriteria of the
quality of the operations, face the stochastic nature
both of the values and
the mechanisms of their
interactions.Thenecessitytodealwiththestochastic
values of all referenced and resulting parameters,
along with their inner transformations, push the
practiceoftheterminaldesigntowardsthesimulation
approach.
Correctuseofthesimulationenablestostudythe
object’s behavior (regarded as the evaluation
of the
parametervaluesovertime),whichistheresultofthe
constituting elements interactions through identified
ties between them. A determinate algorithm of this
Simulation Model of Container Land Terminals
A.L.Kuznetsov,A.V.Kirichenko&J.J.Eglit
A
dmiralMakarovStateUniversityofMaritimeandInnerShipping,St.Petersburg,Russia
ABSTRACT:Thesimulationasatoolforthedesignofportandterminalshasemergedasananswerforthe
demand to enhance the quality and reliability of the project results. Very high costs of the project solution
implementation
and practically total lack of liquidityof transport infrastructure objects always induced the
immensecommercialrisksintheterminalbusiness.Latelytheseriskshavemultipliedsignificantlyduetorapid
changes on the global and regional markets of transport services. Today, many experts come to see this
volatilityasanindicator
ofthenextphaseindevelopmentoftheglobaltradesystemandthederivativecargo
transportationsystem,specificallythestateoftemporalsaturation.Theshiftoftheglobalgoodsvolumesfrom
quickandsteadygrowthtorelativelysmallfluctuationsaroundconstantvaluescausesquickoscillationsinre‐
distribution of demand over
the oversized supply. This new business and economic environment seriously
affected the paradigm of transport terminal design and development techniques. The new operational
environmentofterminalsputarequestforthedesignerstoarrangetheresultsnotintermsof“point”,butin
termsof“functions”.Eventuallyitresultedin
developmentofthemodernobject‐orientedmodelapproach.The
wide spread of this approach witnesses the objective demand for this discipline, while in many aspects it
remainsintheintuitive(pre‐paradigmal)phaseofitsdevelopment.Themainreasonforitisintheproblem
definitionitself,whichusuallyisformulated
asthesimulationofagiventerminal.Atthesametime,thetaskis
toassesstheoperationalcharacteristicsoftheterminalengagedinprocessingofagivencombinationofcargo
flows.Consequently,itisnottheterminalthatshouldbesimulated,buttheprocessesofcargoflowshandling
performed
by this terminal under investigation. Another problem that restricts the practical spread of
simulationisinthemodeladequacy.Amodelwhichadequacyisnotprovedhasnognoseologicalvalueatall.
Thepaperdescribestheapproachaimedatdevelopmentofthemodelswiththefeaturesdiscussedabove.
http://www.transnav.eu
the International Journal
on Marine Navigation
and Safety of Sea Transportation
Volume 12
Number 2
June 2018
DOI:10.12716/1001.12.02.13
322
process gives way to an uncontrolled chance
interactionsoftheobjectsinthe“bullion”ofsoftware
environment selected for the implementation of the
model. The result is the resolving behavior of the
system’s response to the reference inputs and
interfering external influences, studied under
different set of the system state
parameters. The
statistic processing of the received data enables to
obtaintheintegraldistributioncurvesfullydescribing
the parameters as stochastic values. The dynamics
observedinthecourseofthesimulationexperiments
enablestomakethejudgmentsoverthereliabilityof
theestimations.
Modern container terminals, specifically “dry
ports” with
extended functionality and complexity,
byallmeansbelongtothecategoryofobjects,whose
behaviorcannotbeassessedanalytically.
Thisproblemisstudiedbymanyresearchers[3‐5],
reportingimportantandusefulpartialresults.Atthe
sametime,mostoftenthe objectivesofthesestudies
aredeclaredasthesimulationof
theterminals,while
themaingoalistosimulatethecargoflowsontheir
ways trough the terminals. For sea container
terminals these cargo flows and their functional
trajectories are relatively standard, while the wide
specter of functional profiles of dry ports and land
distribution centers causes the wide variety
of the
correspondentcargoflowstructures.Inaddition,the
collection of statistics, input references, experiment
planningandinterpretationofresultsmeetthelackof
the unanimity in the terminology concerning primal
cargo flows and their handling on their routes
throughtheterminal.
This study identifies all possible cargo flow
classes, which
demand different technological
resources in different quantities for their handling.
Forsimulationoftheseflowshandling,ageneralized
universalmodelof‘dryport’typecontainerterminal
is introduced. The structural elements of this model
forms a unified format for the formal descriptionof
technological routs that different classes of cargo
followontheirwaythroughtheterminal.Splittingof
all technological operations into ‘indivisible’ primal
moves provides the possibility of an equally formal
descriptionoftheterminalhandlingsysteminterms
of the equipment used for these operations. The
simplicityofthisdescriptionsignificantlyreducesthe
laboriousnessofthesimulationexperiments,
whichin
itsturnenablestotakeextensivestudiesofthewide
variety of cargo handling systems and large sets of
possiblecargoflows.
Eventually,the discussion ofthe results provides
universal methods of parameter estimations and
recommendationsonutilizationoftheobject‐oriented
simulation as a tool for technological
design of
containerterminals.
2 METHODSANDMATERIALS
2.1 Staticsimulation
Containerizedcargodisplays certainuniquefeatures
unknownin conventional transportation. Breakbulk
could arrive in port loaded in containers, being
registered not in tons, but in teus or boxes, thus
mixing with empty containers. Stripped from
containers, the break bulk
makes the terminal its
generatingpoint,simultaneouslyturningladenboxes
into empties. The container stuffing reverts these
processes. At the same time, both laden and empty
containers are not calculated in tons, as well as the
breakbulkisnotcalculatedinteus.Containerscould
becountedbothinteus
andin‘physicalboxes’.
This ambiguousness of terminology and
interpretationusedfordenotingthemostobviousand
principallyimportant terminal operationsbecomes a
significant problem not only in planning of national
andregionaltransportsystems,analysesofportand
terminal efficiency, but also in their technological
design[6].
Thecargoflows
indryportsbydirectioncouldbe
dividedintoinbound(crossingtheterminalboundary
inwards in any place) and outbound (crossing it
outwards). This is a principal distinguishing feature
between ‘dry’ and ‘sea’ ports, since in the latter the
flows are classified by the direction of crossing the
berthline
(importandexport).
Bythetypeofcargo,bothinboundandoutbound
flowscouldbedividedintobreakbulkandcontainer
flows.
By the mean of processing at the terminal, the
breakbulkandcontainerflowscouldbedividedinto
straight and conversing ones. The straight flows
assumeonlytransportationof
cargothroughterminal
and do not imply the transformation of break bulk
into container and otherwise. The conversing flows
arerepresentedbytwotypes:inboundflowofbreak
bulk which is transferred into outbound container
flow(thestuffingflow),andinbound containerflow
which is transferred into outbound break bulk
flow
(the stripping flow). Loading of break bulk into
containersdemandsfortheflowofemptycontainers
for stuffing. The stripping of laden containers
generates another flow of empty (unloaded)
containers. These to inner terminal flows are called
concomitant.
The volumes of generated and consumed empty
containerscoulddiffer, thusdemanding
thedelivery
fromoutsideincaseofthedeficitanddispatchfrom
theterminalincaseofthesurplus.Inadditiontothe
compensation of this difference, a terminal could
performthefunctionof repositioning ofthe empties
betweenstuffingandstrippingsitesinthehinterland.
Anyway, in addition to
straight (non‐conversed)
flows of break bulk and container, there appears a
separateflowoftheemptycontainers.
Withall these considerations takingintoaccount,
thegeneralizedschemeofcargoflowspassingthedry
portisintroducedinFig.1.
Figure1.Generalschemeofcargoflows.
323
Thefunctionalstructureofthedryportconsistsof
space‐localized processing centers (elements), whose
connections correspond to different terminal cargo‐
handlingoperations,asFig.2shows.
Figure2.Functionalstructureandcargocategories.
In other words, the cargo flows shown by Fig. 1
actually pass over the terminal system’s elements
displayedon Fig. 2.Consequently, every cargo flow
from Fig.1 could be described by the sequence of
functional elements it passes on its technological
route over the terminal, i.e.
12
,, ,
kk k
I
ee e
. In its
turn,each pair of adjacent elements from this list
,
kk
ij
ee
defines one technological operation l=(i,j) of
cargohandling,denotedbyarrowsonFig.2.
Asa result, knowingthe value of a partial cargo
flowQkoveracertaintimeintervalmakesitpossible
to define its technological laboriousness
11 2
; , , ,
kk k k
kI
Qeee e
, as well as to assess the required
operation volume l=(i,j) for any cargo flow k, or
:,
kk
kij
Qee
. The knowledge of the total cargo flow
structure and volume
12
; , , ,
kk k
kI
Qee e
, 1, kK
enables to make a perception of the required
operationvolumesl=(i,j)fortheselectedtimeperiod,
or
1
;,
K
lkk
kij
k
QQee
.
Initsturn,everytechnologicaloperationshownon
Fig.2couldbesplitintoindivisibleʹatom’moves[7],
performedbyoneorseveralspeciesofcargohandling
equipmentclasses(Fig.3).
Every operation assumes its own consequence of
theequipmentspecies
12
,, ,
ll l
L
tt t
,alsoshownonFig.
3. The referenced volume of every operation
l
Q
enablestoassesstherequirementsfortheequipment
involvedinthisoperation
12
; , , ,
lll l
L
Qtt t
.
Summingtogetherthe demands for all groups of
equipment,itispossibletogaintheestimationforthe
required fleet. The same way the relevant
technological resources could be assessed: manning,
fuel consumption, electricity, areas, repair and
maintenancefacilitiesetc.
At the same time, the actions discussed above
describe so
‐called ‘static’ model, using for the
preliminaryestimationsofthetechnologicalresources
overratherlongperiods,e.g.aseasonorayear.
All partia l cargo flows, represented by their
volumes and technological routs, have one equally
important characteristin: the distribution over time
interval. Every flow could be evenly spread over
a
period,condenseinapartofit,overlapwithotheror
fall into empty fragments. Moreover, the
technological resources for partial cargo flows are
restricted,whichcausesthecompetitionoftheflows
forthe resources. In itsturn, it leads to extension of
the cargo flow processing, delays and queues.
This
particularmechanismisresponsibleforthestochastic
fluctuations of the terminal operation parameters,
which rules out the analytical methods of terminal
design.
Figure3.Operationdescriptionintermsofprimalmoves
324
0
200
400
12345678910111213141516171819202122232425262728293031323334353637 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 6 7 686970
ВходЖД
0
200
400
12345678910111213141516171819202122232425262728293031323334353637 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 686970
очер едь навходе ЖД
0
200
400
12345678910111213141516171819202122232425262728293031323334353637 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 5 5 56 57 58 59 60 61 62 63 64 65 66 67 686970
ВыходЖГФ
0
200
400
1 2 3 4 5 6 7 8 9 10 11 12 13 1 4 15 16 17 18 19 20 21 2 2 23 24 25 26 27 28 29 30 31 32 3 3 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 686970
Входоп2
0
500
1000
1500
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 2 4 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 686970
Очередь навходеоп2
0
500
1000
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 2 4 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 686970
Выходоп2
0
500
1000
1 2 3 4 5 6 7 8 9 10111213141516171819202122232425262728293031323334353637 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 686970
Входоп5
0
200
400
600
800
1000
12345678910111213141516171819202122232425262728293031323334353637 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 686970
Очередь навходеоп5
0
500
1000
12345678910111213141516171819202122232425262728293031323334353637 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 6869 70
Выходоп5
0
100
200
12345678910111213141516171819202122232425262728293031323334353637 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 686970
Вход
АМ
0
100
200
300
12345678910111213141516171819202122232425262728293031323334353637 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 686970
очер едь навходе АМ
0
200
400
12345678910111213141516171819202122232425262728293031323334353637 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 686970
ВыходАМ
0
100
200
12345678910111213141516171819202122232425262728293031323334353637 38 39 40 41 42 43 44 45 46 47 48 49 50 5 1 52 53 54 55 56 57 58 59 60 6 1 62 63 64 65 66 67 686970
Входоп6
0
100
200
300
12345678910111213141516171819202122232425262728293031323334353637 38 39 40 41 4 2 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 686970
Очередь навходеоп6
0
100
200
300
12345678910111213141516171819202122232425262728293031323334353637 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 686970
Выходоп6
0
100
200
300
12345678910111213141516171819202122232425262728293031323334353637 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 686970
входоп5
0
200
400
600
800
12345678910111213141516171819202122232425262728293031323334353637 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 6 6 67 686970
Очередь навходеоп5
0
500
1000
12345678910111213141516171819202122232425262728293031323334353637 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 686970
Выходоп5
Figure4.Thescreenshotofthesimulationmonitoring
In order to make judgments over the terminal
parameters interpreted as stochastic values, the
dynamicortheobject‐orientedsimulationapproachis
used.
2.2 Dynamicsimulation
Thedynamicsimulationutilizesthe same functional
modelshownonFig.2.Theprincipaldistinguishing
feature is that this model is used for permanent
re‐
calculation of all values describing the state of the
systemoneverystepof thesimulationprocess. This
statedependsontheinput referenceatthismoment
andthestateofthesystematpreviousmoments.This
statement makes clear that the described paradigm
hereisso‐calleddiscreteevent
simulation[8].
Inordertoexplainthisapproachletusconsideran
example of two cargo flows, whose routes through
the terminal are described by the schemata
1111 1
12576
,, ,,eeeee
and
22222
76521
,, ,,eeeee
. For certainty, let us
assumethat
1
x
e
isthe railcargofront,
2
x
e
isthepre‐
stacking area for rail operations,
5
x
e
is the container
yard,
6
x
e
is the document office and
7
x
e
is the truck
gate to the terminal. The referenced inputs to the
modelarethetrainarrivals(flow1)andtruckarrivals
(flow 2) to the terminal, generated by various
probability patterns. The party size of the arriving
cargo
x
k
q
is also a stocahstic value generated by a
selecteddistribution. Every event underthe selected
notation are ordered lists
111111
12576
; , , , ,
k
qeeeee
and
122222
76521
; , , , ,
k
qeeeee
. An example of this generation is
givenbyTab.1.
At any step of the simulation, the arrival events
putthecorrespondentamountofcargoasajobinthe
queue for handling by an element. If there is a
competitive operational resource allocated for this
elementatthismoment,
itcollectsthecorrespondent
amountofcargoforhandling.Thejobsprocessedby
the element as the server are directed to the next
elementinlinedefinedbythetypeofthecargoflow,
i.e. by the sequences
111111
12576
; , , , ,
k
qeeeee
and
122222
76521
; , , , ,
k
qeeeee
,constitutingthemodellogics.
Table1.Thesampleofinputflowsgeneration
_______________________________________________
Time Input1 Input2
_______________________________________________
1 100 150
2
3 8044
459
5 100 150
6120
7 120 66
8 125 34
9 300 78
10100
11 220 124
12
13 260 67
14 110
15100
16 123 111
17 100 67
18
19 8044
2059
_______________________________________________
All these operations in the model realization are
performed by the standard components – queues,
servers, switches etc. The distribution of the
equipmentpool’recourses
12
,, ,
ll l
L
tt t
byoperations
isdirectedbytheestablishedprioryoftheflowsand
factualdemandsforthem,definedbythesimulation
procedureitself.Thestateofeverycomponentofthe
model reflects the course of operation performance,
relevantdelaysandqueues(Fig.4).
Fig. 5 as an example shows detailed graphics
of
thejobswaitinginthequeuefortransportationfrom
thepre‐stackingareatothecontaineryard.
325
Figure5.Anexampleoftheparameter’sdynamicsrecord
Fig.6 displaysthehistograms ofthisparameter’s
values distributions in the simulation experiments
withdifferentprioritiesoftheresourcesallocation.
Figure6.Queuelengthwithdifferent proprietyassignment
(flow1:flow2)
a)25:75
b)50:50
c)75:20
3 RESULTS
Themechanisms forsimulation ofthe interacting
cargo flows shown in the previous section with the
helpofasimplifiedexampleareimplementedonthe
special object‐oriented software platform. The
interactionofthecargoflowspracticallymeanstheir
competitionfordifferenttechnologicresourcesofthe
terminal:cargo
fronts,warehousefacilities,functional
areas,handlingequipment,personaletc.Forthesame
cargo flows, the sizes of specific resources and the
algorithms of their allocation under deficit led to
different values of the terminal parameters,
responsible for the quality of the rendered
commercialservices.Bythechangingoftechnological
parameters, responsible
for quantitative and
qualitativecharacteristicsoftheterminal,itispossible
to reach the desired level of the service quality,
commonly measured by the length of the queues,
waiting and servicing time. On the other hand, the
owner of the terminal using simulation could build
the perception of the financial
input required to
maintain the quality of servicing. In this way, the
simulationproves outitself to be an efficient tool to
support not only technological, but entrepreneur
decisions.
4 DISCUSSION
The modern highly competitive transport business
environment, with general deterioration of
profitability and growth of cargo flows’ volatility,
makes the
intuitive approaches for taking capital
intensive entrepreneur decisions, based upon
oversimplifiedanalyticalmethods,aseriousthreatfor
theterminalbusiness.Themethodicofthesimulation,
promising today so much for the business, in many
cases comes to realization of a representative
example, to a certain extent reflecting a possible
variant
of a certain demonstrative cargo flow. The
qualityofthemodeloftenisjudgednotbythesizeof
representative set, or accuracy and reliability of the
results,butbythequalityofthegraphicanimationof
the model. Actually, this characteristic is one of the
lessimportantfromgnoseologicalpoint
ofview.The
simulationmodelwhoseadequacyisnotprovedhas
trifling pragmatic value as a tool of design and
decisionsupport.
This postulate is the key stone for the whole
approachdescribedinthispaperandimplementedin
the practically important software toolkit. The most
importantcomponentofany
designprojectbasedon
326
this methodic is a strict, consistent and coherent
provingofthemodel’sadequacytotheprimalobject,
for what the specially designed instruments and
methods are used. Only after a very careful
validation, thorough calibration and sufficient
verification the model could be used as a working
tool.
The utilization of
the carefully adjusted
mechanisms of the object‐oriented simulation,
reached in the course of the adequacy proving
procedure and in the process of rationally planned
movement to the clearly stated goal met every
expectation in several large infrastructural projects
dealing with ‘dry port’ container terminals. The
experiencegainedduring this
studyenabledtoadjust
theperformanceofthisdesignmechanismandrectify
themethodicofitsapplication.
5 CONCLUSIONS
1 Traditional approaches declare the terminal
simulation as their goal, while it is necessary to
simulate the cargo flows passing through the
terminalfunctionalsystem.
2 Seaanddrycontainerterminaldiffer
cardinallyby
the structure, content and functional designation
of their elements, which requires to develop
differentmodelforthem.
3 A generalized universal model structure is
introduced, oriented on awide class of container
terminal of the ‘dry port’ and distribution center
types.
4 Auniversalformatfordescriptionof
typicalcargo
flows passing the terminal along different
technological routes and utilizing its functional
elementsindifferentwayissuggested.
5 A universal format for functional operations
descriptionin terms of the engaged technological
equipmentisintroduced.
6 These universal means of the cargo flows and
description of the required transport
and
technological equipment enable to study various
cargo‐handling systems under different scenarios
easyandefficiently.
7 Since any simulation is only an instrument of
analyses, the procedure of synthesis should be
constructed as a directed changing of technical
parameters, with simulation as a tool for the
comparativestudyof
theirvalues.
8 This controlled search in the range of possible
solutions takes a form of typical scenario
generation, under which the simulation
experiments are planned and performed to
providethestatisticalreliability.
9 Thedescribed techniques are implementedinthe
form of dedicated software with practical
recommendationsontheiruse.
10 The simulation approach for the technological
design discussed in the paper has proved its
worthiness by utilization in several big transport
infrastructureprojects.
REFERENCES
AlexnaderL.Kuznetsov.Genezisagentnogoimitazionnogo
modelirovania v hode razvitia metodov
technologicheskogoproektirovaniaportovIterminalov/
Expluataziavodnogotransporta.N4(58)2009,c.3‐7
Alexnader L. Kuznetsov. Morskaja kontejnernaja
transportno‐technologicheslaya systema/ St.Petersbur:
MANEBprint.house,2017.–320с.
Nevins,M.R.,C.M.Macal,R.Love,andM.
J.Brogen.1998.
Simulation, animation and visualization of seaport
operations.SIMULATION71(2):96‐106.
LegatoP.Asimulationmodellingparadigmfortheoptimal
management of logistics in container terminals. P.
Legato,R. Trunfio // Proceedings of the 21st European
ConferenceonModellingandSimulation.Prague,Czech
Republic.P.479–488.
Najib M. A container terminal management system / M.
Najib,A.ElFazziki,J.Boukachour//Proceedingsofthe
International Conference on Harbour Maritime and
MultimodalLogisticsM&S,2012.P.118–127.
Alexnader L. Kuznetsov. Evoluzya pokazateley
harakterisujushihrabotumorskihportoviterminalov//
VestnikGUMRF.‐2015.‐№6(34).‐C.7‐16.
Alexander
Kuznetsov. Classification of modern container
handlingtechnologies/Cargosystems,June2009,pp.32‐
33
P. Hester, S. Diallo. 2014. “Towards a Theory of Multi ‐
method M&S Approach: Part I.” In Proceedings of the
2014Winter Simulation Conference, edited by A. Tolk,
S.Diallo,I.Ryzhov,L.Yilmaz,S.Buckley,and
J.Miller,
1652‐1663.
