This document summarizes a progress report on a project to develop an optimal short sea shipping (SSS) network model for Korea. Phase 1 was completed in 2005 and involved collecting data on Korea's Pentaport project. Phase 2 in 2006 involved gathering cargo flow data and building an integer programming model to optimize the SSS network. The model aims to minimize transportation costs by determining optimal shipping routes and cargo volumes between ports and cities while considering port capacity and vehicle constraints. Phase 3 in 2007 will involve running the model and analyzing results to identify policy implications for implementing SSS.
1. A Report on
Successful SSS
Models that can
Improve Ports’
Efficiency and
Security while
Reducing
Congestion, Fuel
1
2. Costs, and Pollution
Proposing APEC Economy: U.S.A and Korea
Project number: TPT 02/2006rev1
A Progress Report on the Accomplishment of the Project as of November 1, 2006
The main purpose of this report is to describe what our research team have attempted to
meet the goal of our project and have achieved thus far as of November 1, 2006 and
plan to do in the near future. As described in our proposal, this project is conducted in
three phases: (1) Phase one (2003-2005): Korea provides data on Pentaport project and
organize a conference/workshop on short sea shipping; (2) Phase two (2006): gather
data, build and test short sea shipping flow model; (3) Phase Three (2007): run the
model, analyze results of the model and submit the final report.
Phase one: 2003-2005
The wok of phase one is already completed since the scope of work is to use data from
Pentaport project and organize a conference/workshop on short sea shipping. First of
all, The Pentaport project was completed in 2005 sponsored by Korea’s Ministry of
Maritime Affairs and Fisheries publishing its final report in Korean and providing a
great deal of useful information on how underutilized seaports can be developed
particularly connected with the developments of airport, technoport, business port and
leisure port, therefore, naming this model as Pentaport (five ports in one unit). The
results of the Pentaport project were publicized in various international academic
seminars and also academic journals. Some representing results among these
publications can be summarized as follows:
1. Special Issue: Strategies for Sustainable Coastal Development toward a
Northeast Asian Hub: Comparison between Incheon and Seattle Areas, Korea
Observer, Vol. 34, No. 3 Autumn 2003, Guest-Edited by Peter Rimmer, Young-
Tae Chang and David Fluharty
2
3. 2. Proceedings of Pentaport Seminar to Develop Incheon as Logistics Hub in
Northeast Asia, International Conventional Hall, Jungseok Memorial Library,
Inha University, Incheon, Korea, Oct. 27-29, 2003
3. Proceedings of Northeast Asia Logistics Conference, International
Conventional Hall, Jungseok Memorial Library, Inha University, Incheon,
Korea, April 27-28, 2004
4. Peter Rimmer and Young-Tae Chang, New Roles for Incheon’s Port Authority:
Precedents, Perspectives and Policies, Korea Observer, Vol. 36, No. 2. Summer
2005, pp. 297-322
Secondly, an international seminar on Short Sea Shipping was organized on October
13-14, 2004 at Incheon Hyatt Regency Hotel, Incheon, Korea. The seminar brought
speakers from USA, Korea, Japan, China, Australia, the Netherlands, and Hong Kong.
The details of the presented papers can be referred to from its proceedings as in the
followings (see the cover page of the proceedings and the program from the powerpoint
file presented in Vancouver, August 2006):
Proceedings of Incheon International Seminar on Strategies to Develop Incheon as a
Logistics Hub in Northeast Asia: Short Sea Shipping Approach, Incheon Hyatt Regency
Hotel, Incheon, Korea, Oct. 13-14, 2004
Phase two: 2006
The main goal of the Phase two research is to collect data, build a model and test the
model to find an optimal network model using short sea shipping. To this end, the
research team have tried to collect data from three sources, namely, (1) coastal shipping
data from Korea’s PORTMIS database; (2) national survey data of origin/destination
analysis on Korea’s import and export cargoes and our own research team’s sampled
data; (3) Korea’s Customs Agency Database.
The PORTMIT database provided us with seven years (1997-2004) cargo flows between
28 foreign trade ports and 23 coastal ports in Korea by cargo type, enabling us to
analyze major coastal cargo flows and predict this pattern in the near future. The
3
4. national survey data of import/export provided us with cargo flows between origin and
destination in Korea. Our research team had to collect some additional data by
ourselves to calibrate some of the data that we needed for updating the surveyed data.
(See the powerpoint file that was presented during the APEC Transportation Working
Group in Vancouver, 2006) The Customs Agency Database provided us with detailed
shipmentwise data between Korea’s origin/destination and foreign ports’
origin/destination. The Customs Agency Database is the first output gathered and
published by Korea’s Customs Agency with this much of details. Combining the three
sources of data and adding other socioeconomic data from other various sources, our
research team could grasp all cargo flows between Korea’s inland origin/destination and
foreign ports in terms of the traffic volume, time, cost and mode as well as node. Our
research team had to make our utmost efforts in collecting these data, analyzing huge
volume of data and verifying and validating them.
While collecting the data, our research team have also attempted to build various cargo
flow models to check and test: (1) how current cargo flow network is operating and
what is an optimal solution to minimize the total logistics cost; (2) how the system can
be improved by using more short sea shipping system; and (3) what are the policy
implications to implement short sea shipping for the region. One of the most likely
model that we will use is enclosed in the Appendix of this report for the reference.
To supplement our research work and collect further opinions of other experts, our
research team organized some international seminars and meetings jointly with Korea
Port Economic Association (KPEA) and other academic consortium such as Global U8
Consortium (GU8) (see http://u8.inha.ac.kr) whereby most of our research team
members are active members of the academic societies, for instance, Prof. YT Chang
serves as the President of KPEA and as Secretary-General of GU8. Some of the
intermediate outputs of the data analysis and model building process have been
presented by our research team in the seminars and meetings as follows:
1. Workshop on Input-Output Analysis of Maritime and Ocean Industries organized by
Korea Port Economic Association, June 8, 2006
YT Chang and Paul TW Lee presented two papers
4
5. 2. International Conference on Contemporary Issues of Shipping and Ports in Korea,
organized by Korea Port Economic Association and Jungseok Research Institute of
International Logistics and Trade, Inha University Incheon, Korea, August 8-9, 2006.
YT Chang organized the conference as President of Korea Port Economic Association
where he presented 3 papers and also did discussant job as well.
Paul TW Lee presented papers at Session of Short Sea Shipping and did discussant job
as well.
3. International Conference on National Security, Natural Disasters, Logistics and
Transportation: Assessing the Risks and the Responses, organized by University of
Rhode Island, Kingston, RI, USA, September 25- 26, 2006
YT Chang was Co-chair of the Conference.
Paul TW Lee presented a paper.
4. Winter Conference of Korea Port Economic Association, Seoul, Korea, Dec 4, 2006
YT Chang organized this conference and presented a paper
Paul TW Lee presented two papers
Future plan
Our research team are currently as of writing this report late December, 2006 are
building the final stage of the cargo flow model. We will test and run this model in
January and February, 2007. We plan to present the results of the model during the
APEC TPT meeting in Taiwan, 2007.
Appendix: Flow model of Short Sea Shipping
5
6. An Integer Programming Approach for Short Sea Shipping in Korea
Young-Tae Chang*, Hwa-Joong Kim*, Paul T-W Lee**, Seung-Ho Shin*, Min-Jung Kim*
* Graduate School of Logistics, Inha University, 253 Yonghyeon-dong, Nam-Ku, Incheon, Korea
Corresponding author’s e-mail: hwa-joong.kim@inha.ac.kr
** Department of Logistics and Shipping Management, Kainan University, No.1 Kainan Road, Luzhu,
Taoyuan County 33857, Taiwan
e-mail: paultwlee@inha.ac.kr
Abstract
This study intends to optimize a short sea shipping network in Korea for determining shipping
route, transportation quantities and mode in the network. The problem takes account of two
constraints: maximum cargo volumes capacitated at each seaport and maximum number of
vehicles available at each transportation mode. The objective is to minimize transportation costs
by mode between cargo origin and destination. The problem is formulated as an integer
programming model to solve optimally the problem. The international container cargo data are
used to show the performance of the model. Furthermore, comparative analysis between current
system and the optimal solution derived from our model will suggest several policy
implications.
Keywords: Short sea shipping, Container shipping, Logistics, Integer program.
Topic Area: Shipping, Logistics
6
7. Problem Description
Notations
Parameters
I set of foreign seaports
J set of domestic seaports
K set of domestic cities
M set of modes
cfij cost of shipping an unit of cargo from seaport i to seaport j, which is denoted as
cfij = (h + ctij ) ⋅ ttij + thc j
where h is the inventory holding cost per TEU and unit time, ctij is the transit
cost from per unit TEU and unit time, ttij is the time required for transporting by
ship from seaport i to seaport j, and thcj is the terminal handling charge per unit
TEU
nm TEU that can be carried by mode m
tmjk transportation time via mode m from domestic seaport j to city k or from city k to
domestic seaport j
cdmjk cost required to use one vehicle of mode m to transport from domestic seaport j
to city k or from city k to domestic seaport j, which is denoted as
cd m = ( h ⋅ nm + cmm ) ⋅ t m
jk jk
where cmm is the cost per one vehicle and unit time
cbjb cost required to use one barge to transport from domestic seaport j to domestic
seaport b, which is denoted as
7
8. cb jb = cd 3 + (thc j + thcb ) ⋅ n3
jk
umj available time of mode m at domestic seaport j
umk available time of mode m at city k
sfi supply amount from foreign seaport i
sdk supply amount from city k
dfi demand amount in foreign seaport i
ddk demand amount in city k
aj capacity of domestic seaport j
Decision variables
SIij import amount from foreign seaport i to domestic seaport j
SEji export amount from domestic seaport j to foreign seaport i
DIjk import amount from domestic seaport j to city k
DEkj export amount from city k to domestic seaport j
VImjk number of vehicles of mode m to transport cargos from domestic seaport j to
city/domestic seaport k
VEmkj number of vehicles of mode m to transport cargos from city/domestic seaport k
to domestic seaport j
TIjk import amount excluding transported amount via mode 3 from domestic seaport
j to city k
8
9. TEkj export amount excluding transported amount via mode 3 from city k to domestic
seaport j via domestic seaport b
BIjbk import amount from domestic seaport j to city k via domestic seaport b
BEkbj export amount from city k to domestic seaport j via domestic seaport b
Mixed integer programming model
∑ ∑ cfij ⋅ (SIij + SE ji ) + ∑ ∑ ∑ cd m ⋅ (VI m + VEkj ) + ∑ ∑ cb jb ⋅ (VI 3 + VEbj )
jk jk
m
jb
3
Minimize i∈I j∈J j∈J k∈K m∈M −{3} j∈J b∈J −{ j}
subject to
∑ SI ij = sfi
j∈J for all i ∈ I (1)
∑ DEkj = sd k
j∈J for all k ∈ K (2)
∑ SIij = ∑ DI jk
i∈I k∈K for all j ∈ J (3)
∑ DEkj = ∑ SE ji
k∈K i∈I for all j ∈ J (4)
∑ DI jk = dd k
j∈J for all k ∈ K (5)
∑ SE ji = dfi
j∈J for all i ∈ I (6)
∑ ( SI ij + SE ji ) + ∑ ∑ ( BIbjk + BEkjb ) ≤ a j
i∈I b∈J −{ j} k∈K for all j ∈ J (7)
DI jk + ∑ BIbjk ≤ TI jk + ∑ BI jbk
b∈J −{ j} b∈J −{ j } for all j ∈ J and k ∈ K (8)
9
10. DEkj ≤ TEkj + ∑ BEkbj
b∈J −{ j} for all j ∈ J and k ∈ K (9)
TI jk ≤ ∑ nm ⋅ VI m
for all j ∈ J and k ∈ K
jk
m∈M −{3} (10)
TEkj ≤ ∑ nm ⋅ VEkj
m
m∈M −{3} for all j ∈ J and k ∈ K (11)
∑ BI jbk ≤ n3 ⋅ VI 3
jb
for all j ∈ J and b ∈ J − { j} (12)
k∈K
∑ BEkbj ≤ n3 ⋅ VEbj
3
for all j ∈ J and b ∈ J − { j} (13)
k∈K
∑ t m ⋅ VI m ≤ u m
jk jk j
for all j ∈ J and m ∈ M − {3} (14)
k∈K
∑ tkj ⋅ VEkj ≤ uk
m m m
j∈J for all k ∈ K and m ∈ M − {3} (15)
∑ (t 3 ⋅ VI 3 + t 3 ⋅ VE 3 ) ≤ u 3
for all j ∈ J
jb jb jb jb j
b∈J −{ j } (16)
SI ij , SE ji ≥ 0
for all i ∈ I and j ∈ J (17)
DI jk , DEkj ≥ 0
for all j ∈ J and k ∈ K (18)
TI jk , TEkj ≥ 0
for all j ∈ J and k ∈ K (19)
BI jbk , BEkbj ≥ 0
for all j ∈ J , b ∈ J − { j} , and k ∈ K (20)
VI m , VEkj ≥ 0 and integers
jk
m
for all j ∈ J , k ∈ K , and m ∈ M − {3} (21)
VI 3 , VEbj ≥ 0 and integers
jb
3
for all j ∈ J and b ∈ J − { j} (22)
10