Enhancing Pendulum Nusantara Model in Indonesian Maritime Logistics Network

Abstract

One of the main factors affecting high maritime logistic cost in Indonesia is unbalanced trade between west and east region of Indonesia. Particularly, shipment between two regions were rarely transported using full capacity in both directions. Moreover in many occasion, shipments were carried empty. Therefore, cost of both direction shipment was increased twice or more than normal shipment. Indonesia government has proposed Pendulum Nusantara that guarantees fixed schedule between two regions to cut the logistic cost. This paper generates Indonesia maritime logistic networks using Pendulum Nusantara and enhances it to further bring the cost down and increase profit. The improvements were achieved using combinations of routes that have not considered in Pendulum Nusantara route networks. The problem was modeled as a mixed integer program and a commercial solver was used to generate the solutions. Optimization results show higher profits can be obtained in an acceptable computation time.

Full text

Enhancing Pendulum Nusantara Model in
Indonesian Maritime Logistics Network
Komarudin
System Engineering, Modeling and Simulation (SEMS) Laboratory, Department of Industrial Engineering,
Universitas Indonesia, Jakarta, Indonesia
Email: komarudin@ie.ui.ac.id
Muhammad Reza, Armand Omar Moeis and Arry Rahmawan
System Engineering, Modeling and Simulation (SEMS) Laboratory, Department of Industrial Engineering,
Universitas Indonesia, Jakarta, Indonesia
Email: {muhammad.reza_ti2012@yahoo.com, armand.moeis@ie.ui.ac.id and arry.rahmawan@ie.ui.ac.id}
Abstract— One of the main factors affecting high
maritime logistic cost in Indonesia is unbalanced trade
between west and east region of Indonesia. Particularly,
shipment between two regions were rarely transported
using full capacity in both directions. Moreover in many
occasion, shipments were carried empty. Therefore, cost
of both direction shipment was increased twice or more
than normal shipment. Indonesia government has
proposed Pendulum Nusantara that guarantees fixed
schedule between two regions to cut the logistic cost.
This paper generates Indonesia maritime logistic
networks using Pendulum Nusantara and enhances it to
further bring the cost down and increase profit. The
improvements were achieved using combinations of
routes that have not considered in Pendulum Nusantara
route networks. The problem was modeled as a mixed
integer program and a commercial solver was used to
generate the solutions. Optimization results show higher
profits can be obtained in an acceptable computation
time.
Index Terms— Liner Shipping, Pendulum
Nusantara, Logistic Maritime, Mixed Integer
Programming
I. INTRODUCTION
Sea cargo shipment is an alternative transportation mode
to send goods in between islands [1]. In compared with air
cargo, it is relatively cheaper but with a higher
transportation time. Several commodities that are needed in
an island in Indonesia is sometimes fulfilled using supplies
from other islands. On one hand, excess commodities from
an island can be used to supply other islands. On another
hand, some commodities are not produced within islands so
the commodities need to be brought from outside.
Examples of traded commodities in between Indonesian
islands are cement, rice, coal, automotive, etc.
Commodities trading in between islands and continents
nurtures liner shipping industry [2]. Cargo shipments can
achieve full efficiency if shipments are using full capacities
of the vehicle. Moreover, the accuracies of schedule and
predicted demand can help to increase efficiency.
Based on this situation, Indonesian government has
performed study to generate an efficient liner shipping
network in Indonesia. As a result, Pendulum Nusantara
network has been generated and it is depicted in Figure 1.
Pendulum Nusantara specifies a network that run back and
forth between west to east regions with fixed weekly
schedule. This guarantees that each port is connected to
other ports in Indonesia. In addition, it guarantees regular
shipments can be made because of its fixed schedule.
Recent research suggest that Pendulum Nusantara can be
developed further [3]. Using liner shipping model from
Mulder & Dekker (2014), Van Rijn (2015) and Meijer
(2015) proposed iterated methods to generate router for
liner shipping in Indonesia. In particular, their approaches
aimed to determine the number and type of ships and their
routes that can maximizes weekly profit considering a fixed
weekly demands. The results show that their approaches
result in a significant profit margin from Pendulum
Nusantara.
The current paper aimed to propose number and type of
ships and their routes that can maximizes weekly profit
considering a fixed weekly demands. However, we
consider a higher number of route candidates. Specifically,
routes that have not been considered in previous research.
The rest is organized as follows. Section 2 reviews liner
shipping problem. Section 3 liner shipping mathematical
model due to Mulder & Dekker (2014) is presented.
Section 4 discusses the results of experiments. Finally,
conclusions and further research directions are provided in
Section 5.
Figure 1. Pendulum Nusantara Routes
Maritime logistics networks are main channels for
transporting goods with large volume on long distance.
Three distinctions are made in shipping market: tramp
shipping, industrial shipping and liner shipping. Cargo
117

owners on industrial shipping is also owners of the ships
who strive to minimize the cost of transporting container
between ports. On tramp shipping, vessels are sent to ports
according to availability of container demand. Goods
carried in tramp shipping are bulk cargo. Liner shipping is
the common container shipping type where there are fixed
routes on regular schedules. We are focusing on liner
shipping.
Operation of liner shipping is based on characteristics
associated with routing and scheduling of transporting
containers and cargo. Liner shipper is a company that owns
or operates fleets of container ships. Liner shipping usually
operate on close routes, loading and unloading cargo at any
ports of destination.
The purpose of liner shipping services is to design
network services that can provide a stable and regular
service schedule and also operations that generate profit
(Carranza, 2008). Decision making in liner shipping
consists of three different time-horizon level by Pesenti
(1995): strategic level (3-5 years), tactical level (4-12
months) and operational level (1-4 weeks). Strategic level
has the longest time-horizon. On a strategic level optimal
fleet size is determined. Planning on a tactical level is done
in several months and it involves determining routes uses.
While on operating level that has shortest span of time,
planning allocation of cargo must be done.
Liner shipping company usually operates on various fleet
or various size of vessels on many routes that creates
shipping networks [4] on regular basis, to transport
containers between ports. Liner shipping company is
seeking for optimization technology for an effective cost
planning in operating and enhancing their fleets. This plan
is intended to match capacity of fleets with container
demand effectively. However, in a multi-period planning,
container demand between ports may vary from one period
to another. To cope with container demand pattern from
one period to another, liner shipping company has to adjust
their fleet planning, including fleet size, mix and allocation
of vessels periodically.
II. RESEARCH METHOD
In this research, we try to formulate solutions of strategic,
tactical, and operational planning level of maritime logistic
problem.
On strategic level, the composition of fleet has to be
determined, we call it fleet-design problem. In this research,
it is assumed that company has no fleet in beginning and
the company is sole container shipment provider to fulfill
all of demand.
Constructing network design is the main problem on
tactical planning level. It consists of two problems:
construction of shipping routes and assignment of different
types of ships to routes. For construction of routes, several
types of routing are possible. One can make use of a feeder
network, port-to-port routes and butterfly routes. In this
research, the route that is used is port-to-port.
In case of intra Indonesian shipping, it might be a good
decision to select as hub ports the ports with largest
throughput. The ports used in this research are Belawan,
Tanjung Priok, Tanjung Perak, Banjarmasin, Makassar, and
Sorong. Aggregation of ports are based on throughput and
geographical position of each ports.
Recent research improves the combination of feasible
port routes. Meijer (2015) has designed combination of
three ports on previous research. There were 15
combinations of ship routes and 5 types of vessels that used
to take containers based on PT. Pelindo II (Ports State
Owned Company) that used on this research, therefore total
of routes become 75 combinations. Improvement were
made by enlarging feasible port combination in order to
scaling up possibilities of better solution.
The main problem on operational planning level is
assignment of cargo to ships sailing the determined routes.
This problem is called cargo-routing problem and can be
formulated as an integer linear programming model.
The objective of this research formulation is to optimize
total profit that generated by scenarios. There are two
reasonable scenarios that used by this research. Each
scenario formulates every ports as an origin so that in every
routes, every ships can travel around nearest ports and back
to its origin on final destination. This idea generates 180
new ship routes. The performance is set by summing all
revenue which produced by completing supply-demand
cargo and subtracting all costs which generated by handling
cost, transshipment cost, fuel cost, fixed cost, and port cost.
Mathematical model that is used in this research has
made before [5] with modification of objective function
and few constraints by Meijer (2015). By rewriting the
objective function and some of constraints, model changes
to a Mixed Integer Programming problem and can be used
to determine the optimal fleet, routes and cargo-allocation.
Sets, parameters, decision variables, and equation that
are used in this research, are listed in the following.
Sets:
h
∈
H, Set of ports
t
∈
T
⊆
H, Set of transshipment ports
s
∈
S, Set of ship routes
j
∈
J, Indicator set denoting whether ship
passes both ports and
on ship route , where j = (h1, h2,
s)
k
∈
K, Indicator set denoting whether port
is directly visited after port
on ship route , where k =
(h1, h2, s)
Parameters:
Revenue of transporting one TEU from
port to
Cost of transhipping one TEU in
transshipment port
Cost of (un)loading one TEU in origin
or destination port
Demand with origin port and
destination port
Capacity on ship rute
(0/1) parameter that takes the value 1 if
a ship passes consecutive ports h3
∈
H
and h4
∈
H when sailing from port h1
∈
H to port h2
∈
H on ship route s
∈
S
Fixed cost of using rute
s
∈
S
Distance from sailing from port h1
∈
H
to port h2
∈
H
Fuel price of ship s
∈
S per nautical miles
118

Variables:
Cargo flow on ship route between
consectutive ports and
Integer variable that denotes the number of
times the route is used
Direct cargo flow between ports and
on ship route
Transshipment flow between port and
transshipment port on ship route
Transshipment flow on ship route
between transshipment port and
destination port where the flow to
transshipment port was transported on
ship route
Transshipment flow on ship route
between transshipment port and
transshipment port with destination
port , where the flow to
transshipment port was transported on
route
Objective Function:
Subject to:
The objective function (1) maximizes the profit, which is
equal to revenue minus all costs; fuel costs, transshipment costs,
handling costs and fixed costs. Constraint (2) makes sure that
cargo shipped between every combination of ports does not
exceeding demand for those combinations. Constraint (3) makes
sure that amount of cargo transported on each leg, does not
exceed the capacity of ship sailing this route. Constraint (4)
ensures that all containers which have to be transhipped, will
also be loaded on another route. Constraint (5) defines the
amount of flow between two consecutive ports. Constraint (6)
defines total flow between each two ports in same cycle.
Constraints (7) - (11) all make sure that cargo flow is
nonnegative.
The model was ran using Gurobi solver and Java
programming language. The CPU used in running optimization
model is Intel Core i3 U 380 1.33 GHz.
III. RESULTS AND ANALYSIS
In the experiments, we consider two scenarios. The idea was
to enlarge possibilities of routes combination from Meijer
(2015) default routes. First scenario considers all routes that
each route is a combination of a port and its three closest ports
back and forth so that we generates 180 new routes contains 5
ports combination besides the original 75 routes of Meijer
(2015). In total, we have 255 routes in first scenarios.
Second scenario is same with first scenario, except we only
use two of three closest ports back and forth. On other words,
routes in the second scenarios are routes of first scenarios with
the port next to last port dropped. In this formulation, we have
180 new routes contains 4 ports combination besides the
original 75 routes of Meijer (2015). In total, we have 255
routes in second scenarios.
The results of first and second scenarios are provided in
Table 1 and 2 respectively. They contain selected routes,
quantity and types of ships along with financial calculations.
The routes are represented using sequences of number, with 1
represents Belawan (Medan), 2 represents Batam, 3 represents
Tanjung Priok, 4 represents Surabaya, 5 represents Makasar
and 6 represents Sorong.
However, both scenario are good scenarios of maritime
logistic solution depend on Indonesia government political
will decision. Strategic, tactical and operational level of
decision are covered by each scenario. Until recent times,
there is no decision whether government choose combination
of 5 ports or 4 ports. Other suggestion is that only 4 types of
vessel will be needed, in other hand, the smallest vessel
(Feeder 450) is not needed within scenarios.
Further, we show the performances of two scenarios alto
scnearios provided Bay van Rijn (2015) and Meijer
119

TABLE 1. THE GENERATED ROUTES, TYPES OF SHIPS
RESULTED FROM THE FIRST SCENARIO.
Route
Ship Type
Qty
Fixed Cost
($)
Fuel Cost ($)
Revenue ($)
1-4-2-3-1
Panamax
1750
1
106,256
533,829
2,807,255
4-5-3-2-4
Panamax
1750
1
106,256
217,156
1,890,280
6-5-4-3-6
Panamax
2400
1
148,256
731,604
1,970,690
2-6-2
Feeder 800
1
57,256
392,845
688,000
2-4-2
Panamax
1250
1
78,256
51,467
702,405
3-5-3
Panamax
1250
1
78,256
51,467
934,820
2-3-2
Panamax
2400
1
148,256
79,007
1,927,475
3-4-3
Panamax
2400
1
148,256
79,007
1,915,435
Total
8
871,048
2,
136
,381
12,836,360
TABLE 2. THE GENERATED ROUTES, TYPES OF SHIPS
RESULTED FROM THE SECOND SCENARIO
.
Route
Ship Type
Qty
Fixed Cost
($)
Fuel Cost ($)
Revenue
($)
5-4-3-5
Feeder 800
1
57,256
45,087
811,410
6-4-5-6
Feeder 800
1
57,256
232,687
811,410
4-2-5-4
Panamax
1250
2
156,512
142,753
2,381,985
5-3-2-5
Panamax
1250
1
78,256
141,240
1,039,095
5-3-4-5
Panamax
1750
1
106,256
52,417
1,369,765
3-4-5-3
Panamax
2400
1
148,256
75,393
2,366,935
1-2-1
Panamax
1250
1
78,256
163,175
450,210
5-6-5
Panamax
1750
2
212,512
490,933
2,888,095
Total
10
894,560
1,343,686
12,118,905
TABLE 3. PERFORMANCE OF FIRST
SCENARIO
The first scenario
Revenue
$ 12,836,360
Handling Cost
$ 3,349,624
Transshipment Cost
$ 310,624
Fuel Cost
$ 2,136,381
Fixed Cost
$ 871,048
Port Cost
$ 13,816
Total Profit per Week
$ 6,154,867
Cargo Delivered Percentage
100%
TABLE 4. PERFORMANCE OF SECOND SCENARIO
The second scenario
Revenue
$ 12,118,905
Handling Cost
$ 2,998,395
TransshipmentCost
$
245,208
Fuel Cost
$ 1,343,686
Fixed Cost
$
894,560
Port Cost
$
13,816
Total Profit per Week
$ 6,623,240
Cargo Delivered
Percentage
100%
TABLE 5. PERFORMANCE OF THE BOTH SCENARIOS IN
COMPARED WITH PREVIOUS RESEARCH.
Approaches
Profit ($)
Cargo Delivered
Percentage
Meijer (2015
6,184,308
100%
Van Rijn (2015)
5,364,201
100%
Scenario 1
6,154,867
100%
Scenario 2
6,623,240
100%
These scenario were built based on deterministic supply-
demand data of each ports that also has imbalance trades.
National trade has many factor of improvement which one of
them is maritime logistics. Another thing to be considered is
region development so that each region has an advantage to
balance Indonesia national trade. Future research will be
helpful to suggest government to make strategic decision.
IV. CONCLUSION
In this research, we have provided strategic, tactical, and
operational level of maritime logistic solutions.
We have proposed two scenarios contain a number of ship
routes with ship requirement type and quantity. Further, with
100% cargo delivery we suggest the projected revenue and cost
in every route that can be added to enhance the performance of
Pendulum Nusantara model.
Both scenarios are good solution to Indonesia recent maritime
logistic problem depend on government political will decision.
Future research directions include investigation on the model
on long term decision such as five or ten years in the future.
Moreover, some stochastic variables need to be addresses such
as uncertainty in demands and uncertainty in travel times.
120

REFERENCES
[1] Lawrence, S. (1972). International Sea Transport: The Years
Ahead. Lexington: Lexington Books.
[2] Kalem, H. (2015). Container Liner Shipping Network Design
Using A Path Formulation Model for Indonesia (Master Thesis).
Rotterdam: Erasmus School of Economics, Erasmus University
Rotterdam.
[3] Meijer, J. (2015). Creating A Liner Shipping Network Design
(Bachelor Thesis). Rotterdam: Erasmus University Rotterdam.
[4] van Rijn, L. (2015). Service Network Design for Liner Shipping in
Indonesia (Bachelor Thesis). Rotterdam: Erasmus University
Rotterdam.
[5] Mulder, J., & Dekker, R. (2014). Methods for strategic liner
shipping network design. European Journal of Operational
Research, 235(2), 367–377.
Komarudin, ST, M.Eng, Phd, obtained his
undergraduate degree from Industrial
Engineering department, University of
Indonesia., and master program from
University Teknologi of Malaysia, Phd
Program from Vrije Universiteit Brussel. Mr.
Komarudin is the backbone for the
optimization field in SEMS Lab. Currently,
his research field interests including Lean
and Green Operations in Urban
Transportation, which in line with the
Universitry Research Directives. He is also
interested in the optimization of the current
maritim transport network.
Muhammad Reza, ST. obtained his
undergraduate degree from Industrial
Engineering Universitas Indonesia student
batch 2012 after graduated from 8 Senior
High School Jakarta. He was the President of
Industrial Engineering Student Organization
batch 2014 and President of Faculty of
Engineering Student Organization batch
2015. He has many experiences both in
academic competition, organization and
project committee. As an Industrial
Engineering student, his favorite subjects are
Operation Research, Quality Systems, and
Organization & Psychology of Industry.
Reza is a noisy thinker and choleric-
sanguine guy who believes a lot in
“continuous improvement”.
Armand Omar Moeis, ST, M.Sc, holds a
master degree from Delft University of
Technology, the Netherlands, majoring in
Engineering and Policy Analysis. Prior to his
graduate study, Armand gained his bachelor
degree from Industrial Engineering
Department, University of Indonesia. His
research interests are System Dynamics and
Collaborative Analysis and Learning
whereas his main application areas are
public services and energy. Beside his
position at SEMS, Armand also holds
positions in several business entities. It helps
him to keep up his pace with business and
governmental communities.
Arry Rahmawan, ST, MT, obtained his
undergraduate and master degree from
Industrial Engineering Department,
Universitas Indonesia. Currently, his
research interests are strategic management,
service design & engineering, productivity
improvement, design thinking, simulation
gaming, and system dynamics.
121