Analysis and Simulation of Istanbul Strait Marine Traffic Management Strategies
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
Abstract
Istanbul Strait is one of the most crowded, narrowest straits in the world. More than fifty thousand ships per
year (more than one ship in ten minutes in average) pass through the Istanbul Strait. It is also assumed to be the
most dangerous strait in the world due to its narrowness and nature of the sea flows in the straits. Even though the
number of accidents has decreased in recent years, the risk is still high. Management of Istanbul Strait is indeed a
difficult job required to control many parameters. For example, opening the strait for one or two way traffic,
applying different queuing strategies, or scheduling of limited number of maritime pilots affect the strait traffic. In
order to see the effect of changes in management related strategies, it may not be very practical to really apply the
changes in the traffic management. Simulation of marine traffic in Istanbul Strait is very important to be able to
see the effects of different management strategies and possible changes affecting the marine traffic without really
applying them. This paper presents a simulation model for marine traffic in Istanbul Strait. AutoMod software is
used to develop the simulation model. Different queuing strategies and management related issues are analyzed in
the simulation and their effects are compared in this paper.
Keywords: Marine Traffic Simulation, Marine Traffic Management, Ship Transition Simulation, Istanbul Straits
1. Introduction
27
28
29
30
31
32
33
34
35
The Turkish Straits that include the Istanbul, the Canakkale (Dardanelles) Straits and the Marmara Sea is the only
water route between the Black Sea and the Mediterranean. The Istanbul Strait is about 31 km long, and an average
of 1.5 km wide, with its narrowest place of just 698 meters, passing through the heart of the Istanbul city. Istanbul
has thousands of years of history, is declared as a ‘‘World Heritage City’’ by UNESCO, and also selected to be a
European Capital of Culture by the EU with a population of over 10 million people. As seen from Figure 1, the
strait is located in the heart of Istanbul city.
Figure 1: The Istanbul Strait
The Istanbul Strait is the narrowest waterway in the world. Moreover, three different water flows (i.e., deep,
surface, and inverse flows) make the navigation even more difficult. This narrow, difficult to pass channel is one
of the busiest waterways in the world. Every year, more than 50,000 vessels transit the Bosporus, which makes
the strait four times busier than Panama and three times busier than Suez [9]. The strait is the only waterway for
countries around the Black Sea such as Russia, Ukraine, Bulgaria and Romania (will be called as Black Sea
countries).
TRB 2013 Annual Meeting
Paper revised from original submittal.
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
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
In addition to standard trade from/to the Black Sea countries, oil, natural gas reserves in the Caspian Sea region,
which is estimated to be between 25 and 200 billion barrels respectively [1], increase the traffic density in the
strait. The strait traffic is highly affected from oil transportation, since transportation of oil through Baku-Ceyhan
Pipeline is more expensive than sea transportation (1-2 $ through pipeline, 0.5$ by sea transport, per barrel) [2],
[3]. The importance of the strait for the oil transportation is emphasized in many news and articles [4]-[8]. This
importance is expected to increase in the future with the expected substantial increase (i.e., an increase of 100
million tones of crude oil per year is predicted) in the export of crude oil and liquefied gas from Black Sea
countries [9].
High marine traffic, the strait’s narrowness and water flows make the strait very vulnerable to accidents. For
example, 189 major accidents with many near-casualties have occurred between years 1990-1997. Besides
damaging the environment, in some cases people have died in the accidents. Possible accidents and their damage
to the environment is a very critical issue and studied extensively [10]-[15].
With the given high traffic density, natural difficulties, and consequences of possible accidents, management of
marine traffic in the strait becomes very critical. Many studies and manuals are reported about rules for effective
management of marine traffic in the strait [18]-[25]. Management of marine traffic in Istanbul strait involves
many decision making processes for various parameters and strategies. For example, applying two-way traffic
versus one-way traffic in the strait has advantage and disadvantages. While two-way traffic decreases the ship
waiting time in the queue, it increases the accident risk. On the other hand, maritime pilots travel through the
strait back and forth in the ships in a two-way traffic. In one way traffic, their transportation back to the starting
point in the strait should be supplied by the management using vans or busses on road traffic. Considering the
traffic intensity in Istanbul, especially in rush hours, it is not difficult to see the effect of maritime pilots in the
ship traffic management in the strait. There exist hybrid strategy alternative of one/two way traffic, in which
traffic is two ways in some part of the strait and one-way in the rest. These alternatives should not be applied to
the real case without analyzing it in simulation kind study. Similarly, different queuing techniques used by the
ship traffic management for ship scheduling should be applied to a simulation model, before trying it to the real
system. Thus, it is a great need to simulate marine traffic in Istanbul Strait in order to create an evaluation
environment for possible factors that affect it.
Several studies have been reported for simulation of marine traffic in different straits [26] and Istanbul strait [27],
[28]. In addition to many assumptions, these studies do not report analysis of different management related issues.
These studies mainly focused on the estimation of ship inter-arrival time distributions and analysis of different
distributions. It is also important for a simulation system to consider the rules set by the management. For
example, ship traffic management has set several rules related to geographical properties of different parts of the
strait and ship time.
This paper aims to create a simulation environment for marine traffic in Istanbul strait. This environment will not
only evaluate the effect of different inter-arrival ship arrival distributions, but also help Istanbul Strait Traffic
Management in their decisions by simulating the effects of their decisions on various strategies. Different
strategies and/or parameters such as ship passing time, priority rules, waiting time etc. can be regarded and
maritime traffic can also be investigated in the simulation model. Section 2 gives detailed information about
marine traffic in the strait, section 3 presents the simulation model. Section 4 displays results of different
scenarios implemented in the simulation model, and finally Section 5 concludes the paper.
2. Maritime Traffic in the Strait
TRB 2013 Annual Meeting
Paper revised from original submittal.
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
The passage of vessels through the Turkish Straits is governed by the 1936 Montreux Convention. According to
that, all merchant vessels enjoy, in principle, complete freedom of passage and navigation by day and night,
without being subject to any ‘formalities’ except for sanitary controls and optional towage and pilotage services.
In 1994, Turkey introduced the ‘‘Maritime Traffic Regulations for the Turkish Straits and the Marmara Region”
regarding safety of navigation and the protection of its coastal population and environment. It also established
“Traffic Separation Schemes” (TSS) in the Straits, which follow corresponding International Maritime
Organization (IMO) scheme.
The maritime traffic in the Turkish Straits is exceptionally dense due to the merchant traffic, coasters, fishing
vessels and local traffic causing difficulties with navigation. This dense traffic includes the transport of dangerous
and hazardous cargo (LNG, LPG, oil, chemicals, and other explosive and environmentally hazardous substances).
Internal vessel movement in Istanbul Strait is more than 2000 per day. The passage statistics in the Bosporus
Strait are given in Figure 2 for years 2003-2006. Figure 3 displays the distribution of number of ships throughout
the year 2008. This figure does not include the movement of transiting ships, leisure craft and fishing vessels.
60000
50000
40000
Total number of ships passed
through Istanbul Strait
30000
20000
Number of ships that carry
hazardous and dangerous material
10000
0
2003
100
101
102
2005
2007
2009
2011
103
104
105
106
107
108
Figure 2: Number of ships that passed through the Strait in recent years
Figure 3: Number of ships within a year
Istanbul strait is divided into several sequential regions. Different parts of the strait have different geological
properties. Most important geological parameter is the sharpness of the turn. The Istanbul Strait takes several
sharp turns. At the narrowest point (Kandilli), a 450 course alteration is required. At Yenikoy, the necessary
TRB 2013 Annual Meeting
Paper revised from original submittal.
course alteration is 800. At the turns (Kandilli and Yenikoy) where significant course alterations have to be made,
the rear and forward sights are totally blocked prior to and during the maneuver.
109
110
Criteria
Related to
Sea / Weather
Condition
Vessel-Related
•SP2 Time
•Type / Load
•Length / Speed
•Deep draught
•Pilot request
•Tugboat request
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
Position
•Visibility
•Direction & Intensity of
•Anadolu-Turkeli Pharos line
•Ahirkapi-InciBurnu Pharos line
•Fil-Hamsi Burnu line
•Bogazici Bridge
•Between Kanlica and Vanikoy
Current
Figure 4: Factors affecting the marine traffic
Factors that affect the marine traffic in the strait can be analyzed under three categories: ship, weather, and
position-related factors. Arrival time, ship type, load type, length, deep draught, and pilot and tugboat requests are
vessel-related factors. Visibility, intensity and direction of water flow are the weather-related factors. Positionrelated factors are the restrictions special to the parts of the strait due to narrowness or sharp turns. Figure 4
illustrates these factors. These factors should be considered in the simulation model.
Ships are categorized into six groups based on their size, speed, and load. These groups will be called as SG1,
SG2...SG6 hereafter. In marine traffic management, there are special rules for these groups. For example, when a
SG1 type ship is moving from north to south in some part of the strait (say Kandilli - Yenikoy), a SG1 or SG3
type ship cannot move in the opposite direction. Another example might be that if a SG2 type ship is in the strait,
no SG2 is allowed to enter the strait in the opposite direction until it leaves the strait.
Table 1: Weather related rules
Intensity
of
Current
(IoC)
(Knots)
Genereal
Traffic
Min.
maneu-ver
speed
4≥IoC
Two Way
4<IoC<6
or orkoz
current
6≤IoC<7
or
intense
orkoz
7≤IoC
127
128
129
Vessels carrying dangerous
loads, big vessels (over
200m), deep draught vessels
(over 15m)
Other ships
which can not
provide
minimum
maneuver
speed
That can
attain the
minimum
maneuver
speed
That can not
attain the
minimum
maneuver speed
IoC+4’
Can Enter
Pilot+tug
Pilot+tug
Two Way
10’
Can Enter
Can Not Enter
Pilot+tug
One Way
12’
Can Not Enter
Pilot+tug
Two Way Closed
Weather related factors also affect the marine traffic. Thus, fog or water flow related rules are set by the traffic
management. Examples of these rules are given in Table 1. Position-related factors also affect the marine traffic.
TRB 2013 Annual Meeting
Paper revised from original submittal.
130
131
132
133
134
135
136
137
138
139
For example, 45 degrees of turn is required around Kandilli, which is the narrowest section of the strait with 698
meters. Yenikoy requires 80 degrees of turn. The locations similar to these places include special rules.
There are also general rules to be applied. For example, the distance between two ships cannot be less than 1456
meters (8 gomina), a ship cannot pass another one unless it is necessary (for some locations it is not allowed at
all), or the traffic is restricted to one way when the visibility drops down to less than 1 mile, traffic is closed in
both ways when the visibility drops to less than 0.5 miles. These rules are summarized in Table 2. They may
change depending on three different situations: peace, war, and possibility-of-war times.
Table 2: Examples of Marine Traffic Management Rules
Rule 1
The distance between two ships cannot be less than 1456 meters (8
gomina)
Rule 2
A ship cannot pass another one unless it is necessary
Rule 3
It is not allowed a ship to pass the other even if it is necessary in
some locations
Rule 4
Traffic is restricted to one way when the visibility drops less than 1
mile
Rule 5
Traffic is closed in both ways when the visibility drops less than
0.5 miles
Rule 6
Maximum speed of a ship is 10 sea miles per hour
.
.
.
Rule N
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
.
.
.
When a ship carrying hazardous or dangerous material enters the
strait, no other similar type ship is allowed in the strait
3. Simulation Study
While conducting the simulation study, the steps that are suggested in [29] have been followed. These steps
include:
 Problem Formulation
 Objectives and simulation plan
 Model building
 Data collection
 Distribution fit
 Coding
 Verification and validation
 Experimental design
 Production runs and analysis
At the problem formulation step, first, the problem statement has been made. Istanbul Strait Traffic Management
wants to know the effect of the different strategies when the strait traffic is limited to one way. If the strait is
considered as a set of machines in series with stochastic processing times, then the problem turns out to be
scheduling problem with two different job types that are having different service times and different setup times.
Since the inter-arrival times, processing times and setup times are stochastic; there is no easy analytical solution
to this problem. Moreover, the way the strait is modeled in this research, machines used in series, prevents using
exact analytical models. Therefore simulation is an inevitable tool to analyze this problem.
TRB 2013 Annual Meeting
Paper revised from original submittal.
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
The main objective of the simulation study is to evaluate different strategies when shifting the strait passage route
from one direction to the other. The direction change can only occur if the strait totally cleared. Two main
strategies are identified. One strategy, which is currently used, is the “Fixed Time” strategy. Under this strategy,
the direction of the marine traffic in the strait is changed in predetermined time periods. Currently this time is 12
hours, meaning that, in the first half of the day the vessels are allowed to pass from Marmara to Black Sea while
the other ships that are directed to go from Black sea to Marmara keep waiting.
The second strategy is the “Queue Clearance” strategy. Under this policy, the priority is given to one direction
until the vessels waiting in that direction is totally cleared. To avoid the negative effect of the worst case scenario
for the system, once a priority is given to one direction, new arrivals to the queue can wait for the next turn. Once
the queue is cleared the priority will be given to the other direction and that direction will keep passage right until
it clears its queue. Meanwhile the other queue will be accumulating.
While modeling the system, the strait is divided into 15 sections that can be regarded as machines in series. Each
section can be occupied by only one vessel at a time. This ensures that the minimum distance between vessels is
kept. Vessels arrive from two different resources. Then they move in to their own queues and wait there until they
get warrant to pass the strait. As mentioned before only one of these queues will have passage right at a given
time. Once their passage affirmed and the strait is cleared from the vessels moving in the opposite direction, they
enter the strait. In the strait, a vessel advances from one section to the next section given that the vessel ahead has
left her section. If the vessel ahead is not left her section then the vessels behind slows allowing enough time to
the other one to clear her section. Once a vessel travels through all 15 sections in series, it leaves the system.
To simulate this system AutoMod software and its reports are used (Figure 5). Systematic approach of the
AutoMod is utilized. Simulation model built in AutoMod consists of two major systems components: main
process system and path mover system. In the main process system; loads (vessels), queues, counters, subprocesses, blocks and resources are employed. Different types of loads are defined to represent different types of
vessels.
190
191
Figure 5: AutoMod Runtime View Environment
TRB 2013 Annual Meeting
Paper revised from original submittal.
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
In path mover system two paths are defined. One path starts from Black Sea entrance and goes through the straits
ending at Marmara Sea entrance. The other one extends from Marmara Sea entrance to the Black Sea entrance.
The vessels travel on these paths. These two paths do not overlap. However, very large ships might run over to the
other path at some critical points increasing the risk of having an accident. In addition there is a tunnel
construction under the strait going on currently. Therefore Istanbul Strait Traffic Management changed the traffic
from bi-directional to one way direction. The speed of the vessels and the current in the strait is controlled
through path mover system. The acceleration and deceleration of the vessels are also taken into the consideration
by using stochastic passage times. For each section, minimum passage time, maximum passage time and most
likely passage time information have been used.
Vessel arrival times are taken from Istanbul Strait Traffic Management and grouped based on the vessel type.
Inter arrival times are obtained by considering the times between the successive vessels from the same category.
Since the number of vessels using the strait changes month to month, each inter-arrival times in each month are
analyzed. Then the inter arrival times are fitted to different distributions. For each distribution, after finding the
distribution parameter(s), Chi-square tests are performed to measure goodness of fit. It has been observed that
inter arrival times can be represented by exponential distribution. As an example, the Chi-square test results of the
inter arrival times of the non hazardous ships in one category moving in the direction of Marmara to Black Sea is
shown in Table 3. In this category, in April, 763 vessels arrived to the strait from Marmara direction. The mean
inter arrival time has been calculated 56.55 minutes. For Chi-square test, fifteen data intervals (number of bins)
have been selected. For significance level (Type I error) of =0.05 the critical Chi-square value, 23.6848, is
obtained. Since 9.3526 is less than the critical value, the hypothesis, the inter arrivals fit into the exponential
distribution with mean 56.55 minutes, is accepted.
After finding the appropriate distributions of the data to be used in simulation, coding step took place and
simulation model is built with AutoMod. For verification purposes, the queues and passage of vessels are
animated. The behaviors of the vessels are visually inspected.
Table 3: Statistical Distribution Fitting Results for Interarrival Times (Exponential Dist.)
DataCount:
Number of Bins:
Year:
Month:
Goodness of fit value
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
763
15
2007
4
9.3526
Dist. Parameter
Mean Arrival Time:
Alpha:
Critical Value:
Test Result:
56.55 min
0.05
23.6848
passed
4. Experimentation and Numerical Results
In the objective definition step, it was mentioned above that the performance of the two strait traffic control
policies will be analyzed. There are several performance measurements that can be used. This study mostly
concentrated on the queue statistics, such as average length, maximum length observed, and average waiting time
for each direction queue. When the simulation starts there is a period where the collected statistics is not stable.
After that period, system reaches a state where the critical parameters display a constant behavior. This state is
called steady state. The time required to reach steady state is called warm-up period. It has been observed that as
the system becomes loaded warm up period raises. For fixed time strategy with 12 hour period, the system load is
lower compared to 6-hour period. 6-hour period suggests that it needs 365-days warm-up time whereas 70 days is
sufficient for 12 hour period under fixed time strategy as seen in Figure 6 and Figure 7, respectively. For the
periods between 6 and 12 hours, 300-days warm up period is sufficient as well. So we run the simulation
TRB 2013 Annual Meeting
Paper revised from original submittal.
236
237
238
239
240
241
300+1000 days. First 300 days are used to get the system to steady-sate. No statistics is used in first 300 hundred
days of the simulation.
242
243
244
245
246
247
Figure 6: Warm up analysis when direction change time is 6 hours
Figure 7: Warm up analysis when direction change time is 12 hours
In ‘Fixed Time’ strategy, different fixed times, which is the direction alteration frequency in a day, and their
effects (waiting times and the queue lengths) are analyzed. The results are depicted in Table 4. This result can also
be seen in Figure 8, where the average waiting times for Marmara and Black Sea directions are graphed as well as
TRB 2013 Annual Meeting
Paper revised from original submittal.
248
249
250
251
252
253
254
the total average waiting times. As it can be seen, 8 hours seems to be an optimum time period in an overall
perspective. As a result of this analysis, we recommend Istanbul Strait Traffic Management to change the current
12-hour-fixed time policy to 8-hour fixed time policy. Figure 8 shows there is a linear relationship between
“direction shift period” and “average waiting time” of the vessel after 10 hours of shift periods.
Table 4: Performance of the direction change periods under fixed time strategy
Black Sea
Average
Queue
255
256
5 hours
5.5hours
6 hours
7 hours
8 hours
9 hours
10 hours
11 hours
12 hours
13 hours
14 hours
15 hours
16 hours
17 hours
18 hours
19 hours
20 hours
21 hours
22 hours
23 hours
24 hours
3613.97
1214.68
32.58
16.30
14.18
14.81
15.68
16.54
17.67
18.66
19.76
20.99
22.40
23.57
24.40
26.18
27.17
28.10
29.58
30.63
31.86
TRB 2013 Annual Meeting
Average
Maximum
Waiting
Queue
Time
6762 71990.40
2035 24227.60
139
650.80
102
325.11
63
283.74
65
295.70
58
311.98
61
330.64
72
352.59
69
373.00
66
393.87
73
419.50
75
444.21
84
470.36
88
488.49
88
520.88
87
541.58
92
560.29
101
590.47
99
609.92
105
636.68
Marmara
Average
Queue
76.38
20.51
15.32
12.80
12.90
13.85
14.90
16.03
16.93
17.91
19.38
20.22
21.67
22.73
24.07
25.38
26.50
27.85
28.68
30.12
31.37
Average
Maximum
Waiting
Queue
Time
265
91
84
66
59
51
64
66
63
70
67
74
72
78
81
98
90
95
95
101
108
1525.19
409.59
305.75
256.28
257.73
276.34
297.95
319.56
339.76
358.52
387.07
406.53
432.71
454.00
481.34
507.25
527.43
552.90
573.72
600.21
625.66
Total
Average
Time
36800.21
12326.78
478.1793
290.7529
270.7251
286.0192
304.984
325.094
346.1942
365.7611
390.4731
413.0336
438.4799
462.1822
484.9132
514.0809
534.5005
556.5867
582.1046
605.0668
631.1648
Max No
of Ships
Observed
in
Queues
6768
2052
150
109
77
70
77
79
75
79
76
79
89
87
91
99
92
101
107
107
111
Paper revised from original submittal.
Average Wait Times
700
650
600
550
Minutes
500
450
Average Wait Time
400
Black Sea Avg. Wait
350
Marmara Avg. Wait
300
250
24 hours
23 hours
22 hours
21 hours
20 hours
19 hours
18 hours
17 hours
16 hours
15 hours
14 hours
13 hours
12 hours
11 hours
10 hours
9 hours
8 hours
7 hours
6 hours
200
Direction Shift Periods
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
Figure 8: Average queue waiting times for different periods of direction change
Table 4 and Figure 8 suggest that the average waiting times will exponentially grow when the shift period moves
from 6 hours to the left. In that case, first black sea direction starts accumulating then Marmara direction starts
accumulating. Black sea direction becomes incapable of serving vessels when the period is less than 5.5 hours.
Between 5 and 5.5 hours of periods, Marmara direction is still stable and capable of sending ships through the
strait while Black Sea direction queue keeps accumulating. When the period is less than 5 hours both queues will
keep accumulating ships meaning that they are unstable. So the critical numbers are 5 hours and 5.5 hours for
Marmara and Black sea, respectively.
Figure 9 and Figure 10 give the descriptive statistics of the strait passage times for Marmara and Black Sea
directions respectively. Here, Black Sea direction almost perfectly fits into the normal distribution, while arrivals
from the Marmara direction can also be considered to be normally distributed.
TRB 2013 Annual Meeting
Paper revised from original submittal.
Descriptive Statistics
Variable: Marmara
Anderson-Darling Normality Test
A-Squared:
P-Value:
Mean
StDev
Variance
Skewness
Kurtosis
N
68,5
71,0
73,5
76,0
78,5
81,0
83,5
86,0
88,5
91,0
93,5
96,0
98,5
101,0
103,5
106,0
108,5
111,0
113,5
116,0
118,5
121,0
123,5
126,0
Minimum
1st Quartile
Median
3rd Quartile
Maximum
95% Confidence Interval for Mu
222,813
0,000
95,0972
9,5937
92,0390
0,166159
-5,0E-01
93582
67,370
88,161
94,292
102,093
126,638
95% Confidence Interval for Mu
95,036
94,2
94,7
95,2
9,550
9,637
95% Confidence Interval for Median
95% Confidence Interval for Median
272
273
274
275
276
95,159
95% Confidence Interval for Sigma
94,209
94,369
Figure 9: Descriptive statistics for Marmara entrance
Descriptive Statistics
Variable: BlackSea
Anderson-Darling Normality Test
67,2
69,2
71,2
73,2
75,2
77,2
79,2
81,2
83,2
85,2
87,2
89,2
91,2
93,2
95,2
97,2
99,2
101,2
103,2
105,2
107,2
109,2
111,2
95% Confidence Interval for Mu
A-Squared:
P-Value:
316,491
0,000
Mean
StDev
Variance
Skewness
Kurtosis
N
87,6024
4,7092
22,1768
-8,2E-02
1,14024
94070
Minimum
1st Quartile
Median
3rd Quartile
Maximum
66,736
84,989
87,664
90,393
112,783
95% Confidence Interval for Mu
87,572
87,56
87,61
87,66
87,71
4,688
95% Confidence Interval for Median
277
278
279
87,632
95% Confidence Interval for Sigma
4,731
95% Confidence Interval for Median
87,632
87,698
Figure 10: Descriptive statistics for Black Sea entrance
TRB 2013 Annual Meeting
Paper revised from original submittal.
Strait Passage Times (min)
130
120
110
100
90
80
70
From Black Sea
280
281
282
283
284
285
286
287
288
289
290
From Marmara
Figure 11: Box plot for strait passing times in both directions
The mean and the variance of passing times of vessels coming from Black Sea are much less than the Marmara
sea direction. One might expect shorter queues in Black Sea direction. However it becomes clearer that the queue
length of Black Sea is longer as the system-load gets heavier (i.e. for period less than 8 hours.) This mainly is due
to the delays that occur during changes in passing direction as the vessels wait for the strait to clear from the
vessels that comes from the other direction.
Table 5: Comparison of Queue Clearance and Fixed Time Strategies
Black Sea
Strategy
291
292
293
294
295
296
297
298
299
300
301
302
303
Average Maximum
Queue
Queue
Marmara
Average
Waiting
Time
Total
Average Average
Average Maximum
Waiting Time
Queue
Queue
Time
Max No
of Ships
Observed
in
Queues
Queue
Clearance
11.23
54
224.48
11.45
52
230.26
227.3616
65.00
Optimized
Fixed
Time
14.18
63
283.74
12.90
59
257.73
270.7251
77
After finding the optimum shift time under Fixed Time strategy, the Queue Clearance strategy is evaluated.
Different versions of queue clearance strategy could have been applied. The simplest version can be called Pure
Queue Clearance strategy. Under the pure version, no restriction has been put on the maximum passage right time
for one direction. This type of application is the worst of all the versions of this strategy in terms of waiting time
for a vessel for the worst case scenario. Therefore, ideally, it would be appropriate to put a maximum time on the
passage undertaking to avoid deadlocks in the system and “starving” effects in the queues.
The average time under Pure Queue Clearance strategy has been observed to be less than the waiting time under
the Fixed Time strategy regardless of which direction change period is selected. The maximum queue lengths and
maximum time spent in the queues were acceptable. During 365 days of the simulation, maximum 54 vessels in
southbound queue, 52 vessels on the northbound queue, and 65 vessels in both queues simultaneously have been
TRB 2013 Annual Meeting
Paper revised from original submittal.
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
observed. Therefore no effect of deadlocking or “queue starvation” has been observed in the system. When the
average queue lengths and average queue times are compared, it can be seen that the Queue Clearance strategy
has performed better than Fixed Time strategy with optimum direction change time. As a result, usage of ‘Queue
Clearance’ strategy might be recommended to traffic managements of one way straits, instead of ‘Fixed Time’
strategy. However, in practice, there will be always an upper time limit on the ownership of passage right for any
direction even if the Queue Clearance strategy employed in the strait. Therefore, a good rational strategy is
expected to be at somewhere between ‘Pure Queue Clearance’ strategy and ‘Fixed Time’ strategy.
5. Conclusion
A simulation model of marine traffic in Istanbul Strait is presented in this paper in order to evaluate different
management strategies for Istanbul Strait traffic management. Istanbul Strait is one of the most crowded,
narrowest, and most difficult to pass waterways in the world. Simulation of marine traffic in Istanbul Strait is very
important to be able to see the effects of different management strategies and possible changes in the parameters
that affect the marine traffic. Different traffic management strategies are analyzed in this study. As a result of this
simulation study, we recommend Istanbul Strait Management to apply ‘Queue Clearance’ strategy, in which all
vessels waiting in the queue is cleared before switching the one-way traffic to the opposite direction. We also
recommend switching to 8-hour fixed time strategy from the currently used 12-hour-fixed time strategy if fixed
time strategy is insisted.
REFERENCES
[1] G. Plant, “The Turkish Straits and tanker traffic: an update”, Marine Policy 24 (2000) 193-214
[2] Brito, D. L. (2000), “Congestion of the Turkish Straits: A Market Alternative”, World Congress of the
Econometric Society, Seattle, WA.
[3] E. N. Otay, S. Ozkan, Istanbul Risk Maps, 2005, http://www.ce.boun.edu.tr/otay/SeaAccident/
[4] Lelyveld M. Caspian: pipeline plans becoming commercially less viable.
http://www.rferl.org/nca/features/1999/03, 9 March 1999
[5] Liz Fuller. Kazakh oil through Black Sea. RFE/RL NEWSLINE,vol. 2, Part I, 13 November 1998.
[6] Pannier B. Russia increases transit quota for oil from Kazakhstan. RFE/RL NEWSLINE vol. 3, Part I, 1
March 1999, published on-line at http://www.rferl.org/newsline.
[7] Liz Fuller. Azerbaijani oil reaches Georgian Black Sea terminal.RFE/RL NEWSLINE vol. 3, Part I, 12 March
1999.
[8] Aliriza B. Overview of the Turkish pipeline route alternative. 1 Caspian Crossroads Magazine. Winter 1995:
at http://www.caspcross.com/caspian.
[9] K. Sarioz, A. Kukner and E. Narli, “A Real-Time Ship Maneuvering Simulation Study for the Strait of
Istanbul (Bosporus)”, Journal of Navigation, 1999, vol. 52, no3, pp. 394-410
[10] W. Talley, Vessel damage severity of tanker accidents, Logistics Transport Rev 31(3), 191-207 (1995).
[11] A.L. Kornhauser and W.A. Clark, Quantitative forecast of vessel casualties resulting from additional oil
tanker traffic through the Bosphorus, Report, ALK Associates, Princeton, NJ, September 1995.
[12] E. Anderson and W. Talley, The oil spill size of tanker and barge accidents: Determinants and policy
implications, Land Econ 71, 216-228 (1995).
[13] A.H. Oshins, The risk of oil spills, Risk Management 39(12), 54 (1992)
[14] B. Tan, & E. N. Otay, “Stochastic Modelling and Analysis of Vessel Casualties Resulting from Tanker
Traffic through Narrow Waterways,” Naval Research Logistics, 46[8], (1999), pp. 871-892.
[15] I. Or, E. Kahraman, A simulation study of the accident risk in the istanbul channel, Int. J. Emergency
Manage, (IJEM) (2001).
[16] E. Otay, S. Ozkan, Stochastic prediction of maritime accidents in the Straits of Istanbul, in: Proceedings of
the 3rd International Conference on Oil Spills in the Mediterranean and Black Sea regions, 2003, pp. 92–104.
TRB 2013 Annual Meeting
Paper revised from original submittal.
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
[17] I. Or, M. Sevilir, E. Erkut, An investigation of naval accident probabilities and causes in the Istanbul channel,
J. Manage. Sci. Region. Develop. 1 (2) (1999) 47–60.
[18] OG zman A. (Turkish Under-Secretary of State for Maritime Affairs). Safety in the Black Sea and Passage
through the Bosphorus. Paper to Shipping and Transportation to and from the CIS Conference, London,
March 1995, p. 5.
[19] Toluner S. Rights and duties of Turkey regarding merchant vessels passing through the Straits', in: Turkish
Straits, new problems and solutions, 28, 2829.
[20] Aybay G, Nilufer Oral. Turkey's authority to regulate passage of vessels through the Turkish Straits.
Perceptions. Journal of International Affairs, 1998:3(2), 84.
[21] Regulating traffic flow in the Turkish Straits: a test for modern international law. Emory ILR 1996, available
on-line at http://www.law.emory.edu/EILR, esp. text at nn. 32-38.
[22] Speares S. Bosporus traffic seeks solution. Lloyd's List, 25 July 1998, p. 3.
[23] General Management of Coastal Safety & Salvage Administrations, Turkish Straits Vessel Traffic Service
User’s Guide, Istanbul, 2003.
[24] The IMO rules and recommendations on navigation through the Strait of Istanbul, Strait of Canakkale and
the Marmara Sea, IMO do MSC 63/23, 1994.
[25] The Strait of Istanbul, Sea of Marmara and the Strait of Canakkale Routeing Guide, Turkish Dept. of
Navigation Hydrography and Oceanography, second ed., 1995.
[26] T. Clark, M. Kabil, M. Mostafa, An analysis and simulation of an experimental Suez Canal Traffic Control
System, in: Proceedings of 1983 Winter Simulation Conference, 1983.
[27] E. Kose, E. Basar, E. Demirci, A. Guneroglu, S. Erkebay, Simulation of marine traffic in Istanbul Straits,
Simulation Modelling Practice and Theory 11 (2003) 597–608.
[28] G. Plant, Navigation regime in the Turkish Straits for merchant ships in peacetime, Marine Policy 20 (1996)
15–27.
[29] Banks, Carson, Nelson, and Nicol, Discrete Event System Simulation, p. 16. Prentice-Hall, 3rd ed.
Acknowledgement
This work is supported by the Scientific Research Fund of Fatih University and Istanbul Strait Traffic
Management under the project number P50050901_1”
TRB 2013 Annual Meeting
Paper revised from original submittal.