American International Journal of Contemporary Research Vol. 3 No.

7; July 2013

Application of Queuing Theory to Vehicular Traffic at Signalized Intersection in

Kumasi-Ashanti Region, Ghana

Martin Anokye1
A.R. Abdul-Aziz2
Kwame Annin3
Francis T. Oduro4

Abstract
Traffic congestion is a growing problem in many metropolitan areas as it increases travel time, air pollution,
carbon dioxide (CO2) emissions and fuel use because cars cannot run efficiently. This paper seeks to model the
vehicular traffic flow and explore how vehicular traffic could be minimized using queuing theory in order to
reduce the delays on roads in the Kumasi metropolis of Ghana. The Oforikrom traffic intersection in the Kumasi
metropolis of Ghana is currently operating with one service channel each from the various routes to the
intersection. The results showed that traffic intensity, ρ<1 for all sessions, a condition that suggests a perfect
traffic system. Consequently, smooth flow of traffic was shown since the server at each channel was able to serve
more than the cars in waiting queue when servers resume work. Again, it was found that heavy traffic occurs in
the evening. Stakeholders can task Motor Traffic Transport Unit (MTTU) to check that drivers desist from such
practices so that there will be free flow of traffic in the evening and also promote the use of bikes, which apart
from serving as a form of exercise also helps to reduce fuel consumption thereby saving money for the
Government to tackle problem of other sectors of the economy. Finally, the government of Ghana could introduce
a public transport system so that people do not travel with private cars to their places of work to reduce
congestion on the roads, which in turn boosts productivity.

Keywords: Queuing Theory, Traffic Congestion, Traffic Light System and Transportation
1. Introduction
Queuing theory is the mathematical study of waiting lines, or the act of joining a line (queues). In queueing theory
a model is constructed so that queue lengths and waiting times can be predicted (Sundarapandian, 2009). The
issue of queuing has been a subject of scientific debate for there is no known society that is not confronted with
the problem of queuing. Wherever there is competition for limited resource queuing is likely to occur. The role of
transportation in human life cannot be overemphasized. According to Intikhab et al. (2008), efficient
transportation system plays an important role in catering for the daily necessities in the lives of the citizens. These
include access to amenities and services that are central to the lives of all individuals, like employment, education,
health services and leisure. At the individual level, Wane (2001, p.1) also points out that ‘transportation is a
crucial factor for urban insertion since it gives access to economic activity, facilitates family life and helps in
spinning social networks.
Consequently, in an attempt to acquire vehicles to enable individuals traveling faster and carry out daily
activities easily, cities in the world now witness tremendous motorization during the recent times, especially since
1988 global car population have exceeded 400 million (Walsh, 1990). The reason for this phenomenon, according
to Dimitriou (1991) is that in both the Transportation "currently accounts for 23 percent of the world's
greenhouse-gas emissions," and according to Brad Plumer who writes at the Washington Post “and most of that
unconventional oil is significantly dirtier, from a CO2 perspective, than the traditional stuff. If people in the
developing world keep buying vehicles then simple upgrades in fuel-efficiency alone are not going to be enough
to stop a steady uptick in global temperatures"(Tencer, 2011).

1, 2,3
Lecturer, Mathematics and Statistics Department, Kumasi Polytechnic, Kumasi, Ghana
4
Mathematics Department, Kwame Nkrumah University of Science and Technology. Kumasi, Ghana
23
© Center for Promoting Ideas, USA www.aijcrnet.com

Traffic congestion is a phenomenon of increased disruption of traffic movement on an element of the transport
system. It is most visible when the level of demand for movement approaches or exceeds the present capacity of
the element (Taylor, 1999). As Taylor et al. (2000) argued, traffic congestion presents a common, if not
inevitable, facet of traffic activity in a region, particularly in urban areas. It is also believed that the high volume
of vehicles, the inadequate infrastructure and the irrational distribution of the development are the main reasons
for increasing traffic jam (O’Toole, 2012). The Government in its attempt to salvage the situation has been
expanding road networks in almost all our major cities but the more they expand, the more people import cars for
their domestic use. (Ghanaian Chronicle, 2007).
Quddus et al. (2010) argued that the level of traffic congestion does not affect the severity of road crashes on the
M25 motorway. The impact of traffic flow on the severity of crashes, however, showed an interesting result.
Yannis et al. (2010) developed ordered logit models to provide insight into the human-factors' aspect of the
introduction of advanced technologies with respect to these more sensitive segments of the driver population.
Other studies, Chao et al. (2009), noted that traffic congestion has little or no impact on the frequency of road
accidents on the M25 motorway. Glen et al. (2003) showed how congestion-reduction strategies can induce
additional traffic as a result of economic benefits. Qingyu et al. (2007) suggested a base for the implementation of
urban road congestion pricing and other travel demand management strategies after analyzing the production
mechanism of urban traffic congestion: travelers overlook the negative externality of urban traffic congestion and
then join the congested queues.
Meanwhile, Simulation based analysis on simple network topologies showed that the local de-congestion protocol
can enhance road capacity and prevent congestion collapse in localized settings (Vipin et al., 2012). Jayaram &
Lincoln (2007) identified two polar cases. In one case, the estimation of this number is trivial and therefore
enables further analysis, including the quantification of the relative impact of congestion on distribution costs and
the deduction of empirically testable hypotheses. The other case is considerably less tractable; we consider a
specific instance of this case that was broadly relevant to one of the companies that participated in the research.
Further, Strasbourg et al. (2008) adopted laboratory experiments designed to study the impact of public
information about past departure rates on congestion levels and travel costs. The design was based on a discrete
version of Arnott et al. (1990) bottleneck model. In all treatments, congestion occurs and the observed travel costs
were quite similar to the predicted ones. Subjects' capacity to coordinate was not affected by the availability of
public information on past departure rates, by the number of drivers or by the relative cost of delay. This
seemingly absence of treatment effects was confirmed by finding that a parameter-free reinforcement learning
model best characterizes individual behavior. Luke & Richard (2006) obtain detailed data on road traffic crash
(RTC) casualties, by severity, for each of the eight state and territory jurisdictions in Australia and used these to
estimate and compare the economic impact of RTCs across these regions. The paper makes two fundamental
contributions: (i) it provided a detailed breakdown of estimated RTC casualties, by state and territory regions in
Australia, and (ii) it presented the first sub-national breakdown of RTC costs for Australia.
Kumasi is faced with problem of vehicular traffic congestion which is becoming more serious day after day. Ten
years ago it took about fifteen (15) minutes to travel from Oduom to central market but today it takes about one
hour by car. Accordingly, this article seeks to build a basic model of vehicular traffic based on queuing theory.
Therefore, this will determine the best times for the red, amber and green lights to be either on or off in order to
reduce traffic congestion at the Oforikrom intersection in Kumasi, the Ashanti Region of Ghana.
2. Materials and Method
This section introduces the data sources, discusses the M/M/1/∞ queuing model which this article uses to model
the vehicular traffic flow and to explore how vehicular traffic could be minimized using queuing theory in order
to reduce the delays on roads in Kumasi, capital of Ashanti Region, Ghana
2.1 Data Sources
This paper uses data recorded within the working periods of morning, afternoon and evening where roads are
normally congested at the Oforikrom intersection in Kumasi, Ghana.
The data of traffic from Asokwa to Oforikom was not considered because only negligible size of the traffic used
the route due to road construction being carried at the time of the research.
24
American International Journal of Contemporary Research Vol. 3 No. 7; July 2013

Table 1: Traffic data at Oforikrom Intersection

LOCATION SESSION ARRIVAL SERVICE

Average No. Average
Cars Time(min) No. Cars Time(min)
Tech-to-Oforikrom MORNING 45 1:34 60 1:30
Amakom-to-Oforikrom MORNING 40 1:34 50 1:30
Aboabo-to-Oforikrom MORNING 11 2:30 12 0:34
Tech-to-Oforikrom AFTERNOON 30 1:34 50 1:30
Amakom-to-Oforikrom AFTERNOON 36 1:34 50 1:30
Aboabo-to-Oforikrom AFTERNOON 11 2:30 10 0:34
Tech-to-Oforikrom EVENING 40 1:34 55 1:30
Amakom-to-Oforikrom EVENING 45 1:34 60 1:30
Aboabo-to-Oforikrom EVENING 13 2:30 15 0:34

2.2 M/M/1/ Queueing System

This article will focus on the M/M/1 queuing system. M/M/1 refers to negative exponential arrivals and service
times with a single server. This is the most widely used queuing systems in analysis as pretty much everything is
known about it. M/M/1 is a good approximation for a large number of queuing systems. Suitability of M/M/1
queuing is easy to identify from the server standpoint. For instance a single transmit queue feeding a single link
qualifies as a single server and can be modeled as an M/M/1 queuing system.
2.3 Poisson Arrivals and Process
M/M/1 queuing systems assume a Poisson arrival process. This assumption is a very good approximation for
arrival process in real systems that meet the following rules:
1. The number of customers in the system is very large.
2. The impact of a single customer for the performance of the system is very small, that is, a single customer
consumes a very small percentage of the system resources.
3. All customers are independent, i.e. their decisions to use the system are independent of other users.

Figure 1: M/M/1 Node

2.4 Cars on a Highway

These assumptions are fairly general, so they apply to a large variety of systems. For instance cars entering a
highway could follow these assumptions as follows: Total number of cars driving on the highway is very large.
1. Total number of cars driving on the highway is very large.
2. A single car uses a very small percentage of the highway resources.
3. The decision to enter the highway is independently made by each car driver.
The above observations mean that assuming a Poisson arrival process will be a good approximation of the car
arrivals on the highway. If any one of the three conditions is not met, one cannot assume Poisson arrivals
(eventhelix.com, 2013).

2.5 Probability Density Distribution

25
© Center for Promoting Ideas, USA www.aijcrnet.com

This probability density distribution equation for a Poisson process describes the probability of seeing n arrivals
in a period from 0 to t.

( t ) n   t (1)
Pn (t )  e
n!
Where:
 t is used to define the interval 0 to t
 n is the total number of arrivals in the interval 0 to t.
 λ is the total average arrival rate in arrivals/sec.

2.6 Negative Exponential Arrivals

This equation gives information about how the probability is distributed over a time interval. The probability of
no arrivals taking place over a given interval is obtained by substituting n with 0, and arrives at getting the
following equation:

(2)
P0 (t )  e t
Equation two (2) shows that probability that no arrival takes place during an interval from 0 to t is negative
exponentially related to the length of the interval.
2.7 Poisson Service Times
In an M/M/1 queuing system we assume that service times for customers are also negative exponentially
distributed (i.e. generated by a Poisson process). Unfortunately, this assumption is not as general as the arrival
time distribution. A queuing discipline determines the manner in which the exchange handles calls from
customers. The most common queue discipline is “first- come, first served”, abbreviated as FCFS, in some
inventory applications, the same rule is called “first in- first out” and abbreviated as FIFO.
This principle also serves customers one at a time, however the customer with the shortest waiting time will be
served first.
2.8 Mean Performance Parameters Of The Queue M/M/1/
The equations describing a M/M/1 queuing system are fairly straight forward and easy to use. First we define the
traffic intensity, ρ (sometimes called occupancy). It is defined as the average arrival rate (λ) divided by the
average service rate (µ). For a stable system the average service rate should always be higher than the average
arrival rate. (Otherwise the queues would rapidly race towards infinity). Thus ρ should always be less than one.
Also note that we are talking about average rates here, the instantaneous arrival rate may exceed the service rate.
Over a longer time period, the service rate should always exceed an arrival rate.

 (3)

Mean number of customers in the system (N) :can be found using the following equation

 (1   )  
N   i i  2
 or N  (4)
i 0 (1   ) 1   
Mean number of customers in queue (prior to service)

 2
N q   (i  1) i   (1  (1   ))  (5)
i 0 (1   ) (1   )
From equation five (5) above, as ρ approaches 1 the number of customers would become very large. This can be
easily justified intuitively, ρ will approach 1 when the average arrival rate starts approaching the average service
rate. In this situation, the server would always be busy hence leading to a queue build up (large N).
26
American International Journal of Contemporary Research Vol. 3 No. 7; July 2013
Lastly we obtain the total (mean) waiting time (including the service time):

N  1 1 or 1 (6)
T W     T  Wq 
 (1   )  (1   )    
Mean time spent waiting in queue (prior to service)

 (7)
Tq  W q 
 (1   )
Again we see that as mean arrival rate approaches mean service rate, the waiting time becomes very large
(eventhelix.com, 2013).
3. Results and Discussions
The general objective of this study is to build a basic model of vehicular traffic based on queuing theory and then
use it to determine the best times for the red amber and green lights to be either on or off in order to reduce traffic
congestion at the Oforikrom intersection in Kumasi, in the Ashanti Region Ghana.
3.1 Computation of Queuing Parameters
Computations of parameters to determine the behaviour of traffic were done using the M/M/1 queuing model. It is
assumed that time interval between successive arrivals and serving time is independent and identically distributed.
In any sufficient interval of time at most only one arrival can occur. The system is also assumed to reach a steady
state, a condition that the rate of arrival and service is constant. The queuing discipline observed was first-come-
first-served (FCFS).
The data from Table 1 were used to obtain the mean performance parameters of the M/M/1 queue model as
follows:

Morning Session (Tech-to-Oforikrom)

45
Arrival Rate    34, cars / min
1.34
60
Service Rate    47, cars / min
1.30
34
Traffic Intensity    0.7234
47
1
Mean time spent in the system w   0.0769, min
47(1  0.7234)
0.7234
Mean time spent waiting in queue wq   0.0556, min
47(1  0.7234)
0.7234
Mean number of Cars in the system N   2.610  3
(1  0.7234)
(0.7234) 2
Mean number waiting in queue N q   1.887  2
(1  0.7234)
The same procedure was used to obtain the values from the various routes as shown in Table 2 below.

27
© Center for Promoting Ideas, USA www.aijcrnet.com

Table 2: The Situation of Traffic at Oforikrom Intersection

MEAN MEAN MEAN

NO. OF MEAN NO. TIME TIME
CARS IN OF CARS SPENT SPENT
ARRIVAL SERVICE TRAFFIC THE WAITING IN IN
LOCATION SESSION RATE RATE INTENSITY SYSTEM IN QUEUE SYSTEM QUEUE
Tech Morning 34 47 0.7234 3 2 0.0769 0.0556
Amakom Morning 30 39 0.7692 4 3 0.1111 0.0855
Aboabo Morning 5 36 0.1389 1 1 0.0323 0.0045
Tech Afternoon 23 39 0.5897 2 1 0.0625 0.0369
Amakom Afternoon 27 39 0.6923 3 2 0.0833 0.0577
Aboabo Afternoon 5 33 0.1515 1 1 0.0357 0.0054
Tech Evening 30 43 0.6977 3 2 0.0769 0.0537
Amakom Evening 34 47 0.7234 3 2 0.7234 0.0556
Aboabo Evening 6 45 0.1333 1 1 0.0256 0.0034

All the routes lead to Oforikrom traffic intersection. The content of Table 2 is the results of the analysis of data
recorded in Table 1. They were collected within the working hours from morning, afternoon and evening. The
traffic intersection is currently operating with one service channel each from the various routes to Oforikrom
intersection. It is clear from the results that ρ<1 for all sessions.
4. Conclusion
Critical analysis of the data collected at Oforikrom intersections revealed a smooth flow of traffic and perfect
system, since the server at each channel was able to serve more than the cars in waiting queue when servers
resume work. Moreover, there seems to be heavy traffic in the evening. It was observed that in the evening Motor
traffic transport unit (MTTU) do not work and the drivers of commercial cars take advantage of the situation, park
and offload/on-load passengers at unauthorized places very close to the traffic signals, impeding the flow of
traffic when servers resumed work. This situation is less or non-existent in the morning/afternoon.

Government can task MTTU to check that drivers desist from such practices so that there will be free flow of
traffic in the evening. This is because these practices disrupt the flow and may cause traffic to degenerate into a
disorganized mess. We also suggest that the public transport system should be introduced in Ghana so that people
do not travel with cars to their places of work. Thus, use of public transport will be reducing congestion on the
roads, which in turn boosts productivity. Also, stakeholders can promote the use of bikes, which apart from
serving as a form of exercise, helps to reduce fuel consumption thereby saving money for the Government to
tackle problem of other sectors of the economy.
Acknowledgements
We are grateful to our students who helped us to gather data from the traffic intersection.

28
American International Journal of Contemporary Research Vol. 3 No. 7; July 2013

References
Chao, W. Mohammed, AQ., Stephen, GI. (2009). Impact of Traffic Congestion on Road Safety: A Spatial Analysis of
the M25 Motorway in England. Accident analysis and prevention. Volume: 41, Issue: 4, Publisher: Elsevier
Ltd, Pages: 798-808 PubMed: 19540969
Dimitriou, HT. (1990).Transport Planning for Third World Cities. Editor Routelege, Chapman and Hall, Inc., 1990. 432
pp. + xxii. ISBN 0-415-00448-9
Event Helix, (2013). M/M/1 queueing system. Retrieved January 16, 2013, from
http://www.eventhelix.com/realtimemantra/congestioncontrol/m_m_1_queue.htm#.UQAtcobnnEk
Ghanaian Chronicle (2007). Solving Traffic Congestion in Ghana. Retrieved January 16, 2013, from
http://www.modernghana.com/news/143325/1/solving-traffic-congestion-in-ghana.html
Glen, W., Don V., George T. (2003). Measuring Economic Costs of Urban Traffic Congestion to Business.
Transportation Research Record. Volume: 1839, Issue: 1, Publisher: Trans Res Board, Pages: 98-106 DOI:
10.3141/1839-10
Intikhab, QA., Huapu, L., & Shi Y. (2008). Urban Transportation and Equity: A Case Study of Beijing and Karachi. In
Transportation Research Part A: Policy and Practice, 42: 1
Jayaram, KS., Lincoln, W. (2007). The relative impact of consignee behaviour and road traffic congestion on
distribution costs. Transportation Research Part B: Methodological, Elsevier Ltd. Volume: 41, Issue: 9, Pages:
1033-1049 DOI: 10.1016/j.trb.2007.04.005
Luke, BC., Richard, S. (2006). The economic costs of road traffic crashes: Australia, states and territories. Accident
analysis and prevention. Volume: 38, Issue: 6, Pages: 1087-1093 PubMed: 16797462
Optimum Population Trust. (2009). http://www.optimumpopulation.org/opt.more.traffic.html Accessed on 03.03.2009
Qingyu, L., Zhicai, J., Baofeng, S., Hongfei J. (2007). Method Research on Measuring the External Costs of Urban
Traffic Congestion. Journal of Transportation Systems Engineering and Information Technology. Volume: 7,
Issue: 5, Pages: 9-12 DOI: 10.1016/S1570-6672(07)60035-X
Quddus, MA., Chao, W., Stephen, GI. (2010). Road Traffic Congestion and Crash Severity: Econometric Analysis
Using Ordered Response Models. Journal of Transportation Engineering Volume: 136: Issue Number: 5. Pp
424-435, http://dx.doi.org/10.1061/(ASCE)TE.1943-5436.0000044
Randal, O. (2012). Solving the problem of traffic congestion. Retrieved January 16, 2013, from
http://www.ncpa.org/pub/ba773
Strasbourg, F., Laurent, D., Anthony, Z. (2008). Road Traffic Congestion and Public Information. Journal of Transport
Economics and Policy. Volume: 42, Pp: 43-82
Sundarapandian, V. (2009). 7. Queueing Theory: Probability, Statistics and Queueing Theory. PHI Learning.
ISBN 8120338448
Taylor, MAP. (1999). An Extended Family of Traffic Network Equilibria and its Implications for Land Use and
Transport policies. In Meerman, H., Van Der Voode, E., Winkelmans, W. (Eds.), World Transport Research.
Selected Proceedings from the Eight World Conference on Transport Research, Transport Policy, Elsevier-
Pergamon, Oxford, Vol 4; Pp 29-42.
Taylor, MAP., Woolley, JE., and Zito, R. (2000). Integration of the global positioning system and geographical
information systems for traffic congestion studies. In ThillJ.C. (ed.) Geographic Information Systems in
Transportation Research (1st ed). Elsevier Science Limited, Amsterdam, The Netherlands.
Tencer, Daniel (2011). Number Of Cars Worldwide Surpasses 1 Billion; Can The World Handle This Many Wheels?
Retrieved February 07, 2013 from http://www.huffingtonpost.ca/2011/08/23/car-population_n_934291.html
Vipin Jain, Ashlesh Sharma, Lakshminarayanan Subramanian. (2012). Road Traffic Congestion in the Developing
World. Proceedings of the 2nd ACM Symposium on Computing for Development ACM DEV 12 (2012).
ACM Press, Pages: 10. Retrieved on 14th March 2013 from http://www.mendeley.com/research/road-traffic-
congestion-developing-world/DOI: 10.1145/2160601.2160616
Walsh, M. (1990). Global Trends in Motor Vehicle use and Emissions. In Annual Review of Energy, Volume 15, pp.
217-243
Wane, H.R. (2001). Urban Mobility and Emissions : Towards Accurate Standards for Sub-Saharan Africa. A Research
Program and Results on a Sahelian Case: The District of Bamako. At the Session on Air Pollution, Accra,
Ghana, April 18-20, 2001. Retrieved November 03, 2008
http://www.worldbank.org/transport/utsr/accra/wane.pdf
Yannis, G., Antoniou, C., Vardaki, S., Kanellaidis, G. (2010). Older Drivers' Perception and Acceptance of In-Vehicle
Devices for Traffic Safety and Traffic Efficiency. Journal of Transportation Engineering Volume: 136: Issue
Number: 5. pp 472-479, http://dx.doi.org/10.1061/(ASCE)TE.1943-5436.0000063

29