The document summarizes a study comparing algorithms for solving traffic assignment problems. It classified algorithms as link-based (using link flows), path-based (using path flows), or origin-based (using link flows from origins). It reviewed literature on algorithms like Frank-Wolfe (link-based), path equilibration (path-based), and origin-based algorithm. It chose to implement representative algorithms from each class: Frank-Wolfe, conjugate Frank-Wolfe, bi-conjugate Frank-Wolfe (link-based), path equilibration, gradient projection, projected gradient, improved social pressure (path-based), and Algorithm B (origin-based) to compare their performance on benchmark problems.
Traffic Congestion Prediction using Deep Reinforcement Learning in Vehicular ...IJCNCJournal
In recent years, a new wireless network called vehicular ad-hoc network (VANET), has become a popular research topic. VANET allows communication among vehicles and with roadside units by providing information to each other, such as vehicle velocity, location and direction. In general, when many vehicles likely to use the common route to proceed to the same destination, it can lead to a congested route that should be avoided. It may be better if vehicles are able to predict accurately the traffic congestion and then avoid it. Therefore, in this work, the deep reinforcement learning in VANET to enhance the ability to predict traffic congestion on the roads is proposed. Furthermore, different types of neural networks namely Convolutional Neural Network (CNN), Multilayer Perceptron (MLP) and Long Short-Term Memory (LSTM) are investigated and compared in this deep reinforcement learning model to discover the most effective one. Our proposed method is tested by simulation. The traffic scenarios are created using traffic simulator called Simulation of Urban Mobility (SUMO) before integrating with deep reinforcement learning model. The simulation procedures, as well as the programming used, are described in detail. The performance of our proposed method is evaluated using two metrics; the average travelling time delay and average waiting time delay of vehicles. According to the simulation results, the average travelling time delay and average waiting time delay are gradually improved over the multiple runs, since our proposed method receives feedback from the environment. In addition, the results without and with three different deep learning algorithms, i.e., CNN, MLP and LSTM are compared. It is obvious that the deep reinforcement learning model works effectively when traffic density is neither too high nor too low. In addition, it can be concluded that the effective algorithms for traffic congestion prediction models in descending order are MLP, CNN, and LSTM, respectively.
TRAFFIC CONGESTION PREDICTION USING DEEP REINFORCEMENT LEARNING IN VEHICULAR ...IJCNCJournal
In recent years, a new wireless network called vehicular ad-hoc network (VANET), has become a popular research topic. VANET allows communication among vehicles and with roadside units by providing information to each other, such as vehicle velocity, location and direction. In general, when many vehicles likely to use the common route to proceed to the same destination, it can lead to a congested route that should be avoided. It may be better if vehicles are able to predict accurately the traffic congestion and then avoid it. Therefore, in this work, the deep reinforcement learning in VANET to enhance the ability to predict traffic congestion on the roads is proposed. Furthermore, different types of neural networks namely Convolutional Neural Network (CNN), Multilayer Perceptron (MLP) and Long Short-Term Memory (LSTM) are investigated and compared in this deep reinforcement learning model to discover the most effective one. Our proposed method is tested by simulation. The traffic scenarios are created using traffic simulator called Simulation of Urban Mobility (SUMO) before integrating with deep reinforcement learning model. The simulation procedures, as well as the programming used, are described in detail. The performance of our proposed method is evaluated using two metrics; the average travelling time delay and average waiting time delay of vehicles. According to the simulation results, the average travelling time delay and average waiting time delay are gradually improved over the multiple runs, since our proposed method receives feedback from the environment. In addition, the results without and with three different deep learning algorithms, i.e., CNN, MLP and LSTM are compared. It is obvious that the deep reinforcement learning model works effectively when traffic density is neither too high nor too low. In addition, it can be concluded that the effective algorithms for traffic congestion prediction models in descending order are MLP, CNN, and LSTM, respectively.
Traffic Congestion Prediction using Deep Reinforcement Learning in Vehicular ...IJCNCJournal
In recent years, a new wireless network called vehicular ad-hoc network (VANET), has become a popular research topic. VANET allows communication among vehicles and with roadside units by providing information to each other, such as vehicle velocity, location and direction. In general, when many vehicles likely to use the common route to proceed to the same destination, it can lead to a congested route that should be avoided. It may be better if vehicles are able to predict accurately the traffic congestion and then avoid it. Therefore, in this work, the deep reinforcement learning in VANET to enhance the ability to predict traffic congestion on the roads is proposed. Furthermore, different types of neural networks namely Convolutional Neural Network (CNN), Multilayer Perceptron (MLP) and Long Short-Term Memory (LSTM) are investigated and compared in this deep reinforcement learning model to discover the most effective one. Our proposed method is tested by simulation. The traffic scenarios are created using traffic simulator called Simulation of Urban Mobility (SUMO) before integrating with deep reinforcement learning model. The simulation procedures, as well as the programming used, are described in detail. The performance of our proposed method is evaluated using two metrics; the average travelling time delay and average waiting time delay of vehicles. According to the simulation results, the average travelling time delay and average waiting time delay are gradually improved over the multiple runs, since our proposed method receives feedback from the environment. In addition, the results without and with three different deep learning algorithms, i.e., CNN, MLP and LSTM are compared. It is obvious that the deep reinforcement learning model works effectively when traffic density is neither too high nor too low. In addition, it can be concluded that the effective algorithms for traffic congestion prediction models in descending order are MLP, CNN, and LSTM, respectively.
TRAFFIC CONGESTION PREDICTION USING DEEP REINFORCEMENT LEARNING IN VEHICULAR ...IJCNCJournal
In recent years, a new wireless network called vehicular ad-hoc network (VANET), has become a popular research topic. VANET allows communication among vehicles and with roadside units by providing information to each other, such as vehicle velocity, location and direction. In general, when many vehicles likely to use the common route to proceed to the same destination, it can lead to a congested route that should be avoided. It may be better if vehicles are able to predict accurately the traffic congestion and then avoid it. Therefore, in this work, the deep reinforcement learning in VANET to enhance the ability to predict traffic congestion on the roads is proposed. Furthermore, different types of neural networks namely Convolutional Neural Network (CNN), Multilayer Perceptron (MLP) and Long Short-Term Memory (LSTM) are investigated and compared in this deep reinforcement learning model to discover the most effective one. Our proposed method is tested by simulation. The traffic scenarios are created using traffic simulator called Simulation of Urban Mobility (SUMO) before integrating with deep reinforcement learning model. The simulation procedures, as well as the programming used, are described in detail. The performance of our proposed method is evaluated using two metrics; the average travelling time delay and average waiting time delay of vehicles. According to the simulation results, the average travelling time delay and average waiting time delay are gradually improved over the multiple runs, since our proposed method receives feedback from the environment. In addition, the results without and with three different deep learning algorithms, i.e., CNN, MLP and LSTM are compared. It is obvious that the deep reinforcement learning model works effectively when traffic density is neither too high nor too low. In addition, it can be concluded that the effective algorithms for traffic congestion prediction models in descending order are MLP, CNN, and LSTM, respectively.
1
Intermodal Autonomous Mobility-on-Demand
Mauro Salazar1,2, Nicolas Lanzetti1,2, Federico Rossi2, Maximilian Schiffer2,3, and Marco Pavone2
Abstract—In this paper we study models and coordination poli-
cies for intermodal Autonomous Mobility-on-Demand (AMoD),
wherein a fleet of self-driving vehicles provides on-demand
mobility jointly with public transit. Specifically, we first present
a network flow model for intermodal AMoD, where we capture
the coupling between AMoD and public transit and the goal is
to maximize social welfare. Second, leveraging such a model,
we design a pricing and tolling scheme that allows the system
to recover a social optimum under the assumption of a perfect
market with selfish agents. Third, we present real-world case
studies for the transportation networks of New York City and
Berlin, which allow us to quantify the general benefits of
intermodal AMoD, as well as the societal impact of different
vehicles. In particular, we show that vehicle size and powertrain
type heavily affect intermodal routing decisions and, thus, system
efficiency. Our studies reveal that the cooperation between AMoD
fleets and public transit can yield significant benefits compared
to an AMoD system operating in isolation, whilst our proposed
tolling policies appear to be in line with recent discussions for
the case of New York City.
I. INTRODUCTION
TRAFFIC congestion is soaring all around the world. Besidesmere discomfort for passengers, congestion causes severe
economic and environmental harm, e.g., due to the loss of
working hours and pollutant emissions such as CO2, partic-
ulate matter, and NOx [1]. In 2013, traffic congestion cost
U.S. citizens 124 Billion USD [2]. Notably, transportation
remains one of a few sectors in which emissions are still
increasing [3]. Governments and municipalities are struggling
to find sustainable ways of transportation that can match
mobility needs and reduce environmental harm as well as
congestion.
To achieve sustainable modes of transportation, new mobil-
ity concepts and technology changes are necessary. However,
the potential to realize such concepts in urban environments is
limited, since upgrades to available infrastructures (e.g., roads
and subway lines) and their capacity are often extremely costly
and require decades-long planning timelines. Thus, mobility
concepts that use existing infrastructure in a more efficient way
are especially attractive. In this course, mobility-on-demand
services appear to be particularly promising. Herein, two main
concepts exist. On the one hand, free floating car sharing
systems strive to reduce the total number of private vehicles
in city centers. However, these systems offer limited flexibility
and are generally characterized by low adoption rates that
result from low vehicle availabilities due to the difficulty of
1Institute for Dynamic Systems and Control ETH Zürich, Zurich (ZH),
Switzerland {samauro,lnicolas}@ethz.ch
2Department of Aeronautics and Astro.
A study and implementation of the transit route network design problem for a ...csandit
The design of public transportation networks presup
poses solving optimization problems,
involving various parameters such as the proper mat
hematical description of networks, the
algorithmic approach to apply, and also the conside
ration of real-world, practical
characteristics such as the types of vehicles in th
e network, the frequencies of routes, demand,
possible limitations of route capacities, travel de
cisions made by passengers, the environmental
footprint of the system, the available bus technolo
gies, besides others. The current paper
presents the progress of the work that aims to stud
y the design of a municipal public
transportation system that employs middleware techn
ologies and geographic information
services in order to produce practical, realistic r
esults. The system employs novel optimization
approaches such as the particle swarm algorithms an
d also considers various environmental
parameters such as the use of electric vehicles and
the emissions of conventional ones.
A STUDY AND IMPLEMENTATION OF THE TRANSIT ROUTE NETWORK DESIGN PROBLEM FOR A ...cscpconf
The design of public transportation networks presupposes solving optimization problems,
involving various parameters such as the proper mathematical description of networks, the
algorithmic approach to apply, and also the consideration of real-world, practical
characteristics such as the types of vehicles in the network, the frequencies of routes, demand,
possible limitations of route capacities, travel decisions made by passengers, the environmental
footprint of the system, the available bus technologies, besides others. The current paper
presents the progress of the work that aims to study the design of a municipal public
transportation system that employs middleware technologies and geographic information
services in order to produce practical, realistic results. The system employs novel optimization
approaches such as the particle swarm algorithms and also considers various environmental
parameters such as the use of electric vehicles and the emissions of conventional ones.
Conflict-free dynamic route multi-agv using dijkstra Floyd-warshall hybrid a...IJECEIAES
Autonomous Guided Vehicle is a mobile robot that can move autonomously on a route or lane in an indoor or outdoor environment while performing a series of tasks. Determination of the shortest route on an autonomous guided vehicle is one of the optimization problems in handling conflict- free routes that have an influence on the distribution of goods in the manufacturing industry's warehouse. Pickup and delivery processes in the distribution on AGV goods such as scheduling, shipping, and determining the route of vehicle with short mileage characteristics, is very possible to do simulations with three AGV units. There is a windows time limit on workstations that limits shipping. The problem of determining the route in this study is considered necessary as a multi-vehicle route problem with a time window. This study aims to describe the combination of algorithms written based on dynamic programming to overcome the problem of conflict-free AGV routes using time windows. The combined approach of the Dijkstra and Floyd-Warshall algorithm results in the optimization of the closest distance in overcoming conflict-free routes.
International Journal of Business and Management Invention (IJBMI)inventionjournals
International Journal of Business and Management Invention (IJBMI) is an international journal intended for professionals and researchers in all fields of Business and Management. IJBMI publishes research articles and reviews within the whole field Business and Management, new teaching methods, assessment, validation and the impact of new technologies and it will continue to provide information on the latest trends and developments in this ever-expanding subject. The publications of papers are selected through double peer reviewed to ensure originality, relevance, and readability. The articles published in our journal can be accessed online
Fuzzy Logic Model for Traffic CongestionIOSR Journals
Abstract: Traffic congestion has become a serious problem in the urban districts. This is mainly due to the
rapid increase in the number and the use of vehicles. Travel time, travel safety, environmental quality, and life
quality are all adversely affected by traffic congestion. Many traffic control systems have been developed and
installed to alleviate the problem with limited success. Traffic demands are still high and increasing. The main
focus of this report is to introduce a versatile fuzzy logic traffic flow model capable of making optimal traffic
predictions. This model can be used to evaluate various traffic-light timing plans. More importantly, it provides
a framework for implementing adaptive traffic signal controllers based on fuzzy logic technology. When
implemented it solved the problem of waiting time, travel cost, accident, traffic congestion.
Key words: Traffic Congestion, fuzzy logic, Traffic Density, fuzzy controller, conventional controller.
A New Paradigm in User Equilibrium-Application in Managed Lane PricingCSCJournals
Ineffective use of the High-Occupancy-Vehicle (HOV) lanes has the potential to decrease the overall roadway throughput during peak periods. Excess capacity in HOV lanes during peak periods can be made available to other types of vehicles, including single occupancy vehicles (SOV) for a price (toll). Such dual use lanes are known as “Managed Lanes.” The main purpose of this research is to propose a new paradigm in user equilibrium to predict the travel demand for determining the optimal fare policy for managed lane facilities. Depending on their value of time, motorists may choose to travel on Managed Lanes (ML) or General Purpose Lanes (GPL). In this study, the features in the software called Toll Pricing Modeler version 4.3 (TPM-4.3) are described. TPM-4.3 is developed based on this new user equilibrium concept and utilizes it to examine various operating scenarios. The software has two built-in operating objective options: 1) what would the ML operating speed be for a specified SOV toll, or 2) what should the SOV toll be for a desired minimum ML operating speed. A number of pricing policy scenarios are developed and examined on the proposed managed lane segment on Interstate 30 (I-30) in Grand Prairie, Texas. The software provides quantitative estimates of various factors including toll revenue, emissions and system performance such as person movement and traffic speed on managed and general purpose lanes. Overall, among the scenarios examined, higher toll rates tend to generate higher toll revenues, reduce overall CO and NOx emissions, and shift demand to general purpose lanes. On the other hand, HOV preferential treatments at any given toll level tend to reduce toll revenue, have no impact on or reduce system performance on managed lanes, and increase CO and NOx emissions.
Modeling business management systems transportationSherin El-Rashied
Introduction
How IT &Business Process Fit Together
What is modeling?
What is Simulation?
Modeling & Simulation in Business Process Management
The Seven-Step Model-Building Process
Transportation
An overview on transportation modeling
Transport model scope & structure
Car Traffic Jam Problem
Aim of Transportation Model
Types of Traffic Models
Microscopic Traffic model & Simulation
Cellular Automaton model
Conclusion
Solving Transportation Problem by Software Application
Class Example
Vehicular ad hoc networks (VANETs) have seen tremendous growth in the last decade, providing a vast
range of applications in both military and civilian activities. The temporary connectivity in the vehicles can also
increase the driver’s capability on the road. However, such applications require heavy data packets to be shared on
the same spectrum without the requirement of excessive radios. Thus, e-client approaches are required which can
provide improved data dissemination along with the better quality of services to allow heavy traffic to be easily
shared between the vehicles. In this paper, an e-client data dissemination approach is proposed which not only
improves the vehicle to vehicle connectivity but also improves the QoS between the source and the destination. The
proposed approach is analyzed and compared with the existing state-of-the-art approaches. The effectiveness of the
proposed approach is demonstrated in terms of the significant gains attained in the parameters namely, end to end
delay, packet delivery ratio, route acquisition time, throughput, and message dissemination rate in comparison with
the existing approaches.
1
Intermodal Autonomous Mobility-on-Demand
Mauro Salazar1,2, Nicolas Lanzetti1,2, Federico Rossi2, Maximilian Schiffer2,3, and Marco Pavone2
Abstract—In this paper we study models and coordination poli-
cies for intermodal Autonomous Mobility-on-Demand (AMoD),
wherein a fleet of self-driving vehicles provides on-demand
mobility jointly with public transit. Specifically, we first present
a network flow model for intermodal AMoD, where we capture
the coupling between AMoD and public transit and the goal is
to maximize social welfare. Second, leveraging such a model,
we design a pricing and tolling scheme that allows the system
to recover a social optimum under the assumption of a perfect
market with selfish agents. Third, we present real-world case
studies for the transportation networks of New York City and
Berlin, which allow us to quantify the general benefits of
intermodal AMoD, as well as the societal impact of different
vehicles. In particular, we show that vehicle size and powertrain
type heavily affect intermodal routing decisions and, thus, system
efficiency. Our studies reveal that the cooperation between AMoD
fleets and public transit can yield significant benefits compared
to an AMoD system operating in isolation, whilst our proposed
tolling policies appear to be in line with recent discussions for
the case of New York City.
I. INTRODUCTION
TRAFFIC congestion is soaring all around the world. Besidesmere discomfort for passengers, congestion causes severe
economic and environmental harm, e.g., due to the loss of
working hours and pollutant emissions such as CO2, partic-
ulate matter, and NOx [1]. In 2013, traffic congestion cost
U.S. citizens 124 Billion USD [2]. Notably, transportation
remains one of a few sectors in which emissions are still
increasing [3]. Governments and municipalities are struggling
to find sustainable ways of transportation that can match
mobility needs and reduce environmental harm as well as
congestion.
To achieve sustainable modes of transportation, new mobil-
ity concepts and technology changes are necessary. However,
the potential to realize such concepts in urban environments is
limited, since upgrades to available infrastructures (e.g., roads
and subway lines) and their capacity are often extremely costly
and require decades-long planning timelines. Thus, mobility
concepts that use existing infrastructure in a more efficient way
are especially attractive. In this course, mobility-on-demand
services appear to be particularly promising. Herein, two main
concepts exist. On the one hand, free floating car sharing
systems strive to reduce the total number of private vehicles
in city centers. However, these systems offer limited flexibility
and are generally characterized by low adoption rates that
result from low vehicle availabilities due to the difficulty of
1Institute for Dynamic Systems and Control ETH Zürich, Zurich (ZH),
Switzerland {samauro,lnicolas}@ethz.ch
2Department of Aeronautics and Astro.
A study and implementation of the transit route network design problem for a ...csandit
The design of public transportation networks presup
poses solving optimization problems,
involving various parameters such as the proper mat
hematical description of networks, the
algorithmic approach to apply, and also the conside
ration of real-world, practical
characteristics such as the types of vehicles in th
e network, the frequencies of routes, demand,
possible limitations of route capacities, travel de
cisions made by passengers, the environmental
footprint of the system, the available bus technolo
gies, besides others. The current paper
presents the progress of the work that aims to stud
y the design of a municipal public
transportation system that employs middleware techn
ologies and geographic information
services in order to produce practical, realistic r
esults. The system employs novel optimization
approaches such as the particle swarm algorithms an
d also considers various environmental
parameters such as the use of electric vehicles and
the emissions of conventional ones.
A STUDY AND IMPLEMENTATION OF THE TRANSIT ROUTE NETWORK DESIGN PROBLEM FOR A ...cscpconf
The design of public transportation networks presupposes solving optimization problems,
involving various parameters such as the proper mathematical description of networks, the
algorithmic approach to apply, and also the consideration of real-world, practical
characteristics such as the types of vehicles in the network, the frequencies of routes, demand,
possible limitations of route capacities, travel decisions made by passengers, the environmental
footprint of the system, the available bus technologies, besides others. The current paper
presents the progress of the work that aims to study the design of a municipal public
transportation system that employs middleware technologies and geographic information
services in order to produce practical, realistic results. The system employs novel optimization
approaches such as the particle swarm algorithms and also considers various environmental
parameters such as the use of electric vehicles and the emissions of conventional ones.
Conflict-free dynamic route multi-agv using dijkstra Floyd-warshall hybrid a...IJECEIAES
Autonomous Guided Vehicle is a mobile robot that can move autonomously on a route or lane in an indoor or outdoor environment while performing a series of tasks. Determination of the shortest route on an autonomous guided vehicle is one of the optimization problems in handling conflict- free routes that have an influence on the distribution of goods in the manufacturing industry's warehouse. Pickup and delivery processes in the distribution on AGV goods such as scheduling, shipping, and determining the route of vehicle with short mileage characteristics, is very possible to do simulations with three AGV units. There is a windows time limit on workstations that limits shipping. The problem of determining the route in this study is considered necessary as a multi-vehicle route problem with a time window. This study aims to describe the combination of algorithms written based on dynamic programming to overcome the problem of conflict-free AGV routes using time windows. The combined approach of the Dijkstra and Floyd-Warshall algorithm results in the optimization of the closest distance in overcoming conflict-free routes.
International Journal of Business and Management Invention (IJBMI)inventionjournals
International Journal of Business and Management Invention (IJBMI) is an international journal intended for professionals and researchers in all fields of Business and Management. IJBMI publishes research articles and reviews within the whole field Business and Management, new teaching methods, assessment, validation and the impact of new technologies and it will continue to provide information on the latest trends and developments in this ever-expanding subject. The publications of papers are selected through double peer reviewed to ensure originality, relevance, and readability. The articles published in our journal can be accessed online
Fuzzy Logic Model for Traffic CongestionIOSR Journals
Abstract: Traffic congestion has become a serious problem in the urban districts. This is mainly due to the
rapid increase in the number and the use of vehicles. Travel time, travel safety, environmental quality, and life
quality are all adversely affected by traffic congestion. Many traffic control systems have been developed and
installed to alleviate the problem with limited success. Traffic demands are still high and increasing. The main
focus of this report is to introduce a versatile fuzzy logic traffic flow model capable of making optimal traffic
predictions. This model can be used to evaluate various traffic-light timing plans. More importantly, it provides
a framework for implementing adaptive traffic signal controllers based on fuzzy logic technology. When
implemented it solved the problem of waiting time, travel cost, accident, traffic congestion.
Key words: Traffic Congestion, fuzzy logic, Traffic Density, fuzzy controller, conventional controller.
A New Paradigm in User Equilibrium-Application in Managed Lane PricingCSCJournals
Ineffective use of the High-Occupancy-Vehicle (HOV) lanes has the potential to decrease the overall roadway throughput during peak periods. Excess capacity in HOV lanes during peak periods can be made available to other types of vehicles, including single occupancy vehicles (SOV) for a price (toll). Such dual use lanes are known as “Managed Lanes.” The main purpose of this research is to propose a new paradigm in user equilibrium to predict the travel demand for determining the optimal fare policy for managed lane facilities. Depending on their value of time, motorists may choose to travel on Managed Lanes (ML) or General Purpose Lanes (GPL). In this study, the features in the software called Toll Pricing Modeler version 4.3 (TPM-4.3) are described. TPM-4.3 is developed based on this new user equilibrium concept and utilizes it to examine various operating scenarios. The software has two built-in operating objective options: 1) what would the ML operating speed be for a specified SOV toll, or 2) what should the SOV toll be for a desired minimum ML operating speed. A number of pricing policy scenarios are developed and examined on the proposed managed lane segment on Interstate 30 (I-30) in Grand Prairie, Texas. The software provides quantitative estimates of various factors including toll revenue, emissions and system performance such as person movement and traffic speed on managed and general purpose lanes. Overall, among the scenarios examined, higher toll rates tend to generate higher toll revenues, reduce overall CO and NOx emissions, and shift demand to general purpose lanes. On the other hand, HOV preferential treatments at any given toll level tend to reduce toll revenue, have no impact on or reduce system performance on managed lanes, and increase CO and NOx emissions.
Modeling business management systems transportationSherin El-Rashied
Introduction
How IT &Business Process Fit Together
What is modeling?
What is Simulation?
Modeling & Simulation in Business Process Management
The Seven-Step Model-Building Process
Transportation
An overview on transportation modeling
Transport model scope & structure
Car Traffic Jam Problem
Aim of Transportation Model
Types of Traffic Models
Microscopic Traffic model & Simulation
Cellular Automaton model
Conclusion
Solving Transportation Problem by Software Application
Class Example
Vehicular ad hoc networks (VANETs) have seen tremendous growth in the last decade, providing a vast
range of applications in both military and civilian activities. The temporary connectivity in the vehicles can also
increase the driver’s capability on the road. However, such applications require heavy data packets to be shared on
the same spectrum without the requirement of excessive radios. Thus, e-client approaches are required which can
provide improved data dissemination along with the better quality of services to allow heavy traffic to be easily
shared between the vehicles. In this paper, an e-client data dissemination approach is proposed which not only
improves the vehicle to vehicle connectivity but also improves the QoS between the source and the destination. The
proposed approach is analyzed and compared with the existing state-of-the-art approaches. The effectiveness of the
proposed approach is demonstrated in terms of the significant gains attained in the parameters namely, end to end
delay, packet delivery ratio, route acquisition time, throughput, and message dissemination rate in comparison with
the existing approaches.
Embracing GenAI - A Strategic ImperativePeter Windle
Artificial Intelligence (AI) technologies such as Generative AI, Image Generators and Large Language Models have had a dramatic impact on teaching, learning and assessment over the past 18 months. The most immediate threat AI posed was to Academic Integrity with Higher Education Institutes (HEIs) focusing their efforts on combating the use of GenAI in assessment. Guidelines were developed for staff and students, policies put in place too. Innovative educators have forged paths in the use of Generative AI for teaching, learning and assessments leading to pockets of transformation springing up across HEIs, often with little or no top-down guidance, support or direction.
This Gasta posits a strategic approach to integrating AI into HEIs to prepare staff, students and the curriculum for an evolving world and workplace. We will highlight the advantages of working with these technologies beyond the realm of teaching, learning and assessment by considering prompt engineering skills, industry impact, curriculum changes, and the need for staff upskilling. In contrast, not engaging strategically with Generative AI poses risks, including falling behind peers, missed opportunities and failing to ensure our graduates remain employable. The rapid evolution of AI technologies necessitates a proactive and strategic approach if we are to remain relevant.
Introduction to AI for Nonprofits with Tapp NetworkTechSoup
Dive into the world of AI! Experts Jon Hill and Tareq Monaur will guide you through AI's role in enhancing nonprofit websites and basic marketing strategies, making it easy to understand and apply.
Biological screening of herbal drugs: Introduction and Need for
Phyto-Pharmacological Screening, New Strategies for evaluating
Natural Products, In vitro evaluation techniques for Antioxidants, Antimicrobial and Anticancer drugs. In vivo evaluation techniques
for Anti-inflammatory, Antiulcer, Anticancer, Wound healing, Antidiabetic, Hepatoprotective, Cardio protective, Diuretics and
Antifertility, Toxicity studies as per OECD guidelines
June 3, 2024 Anti-Semitism Letter Sent to MIT President Kornbluth and MIT Cor...Levi Shapiro
Letter from the Congress of the United States regarding Anti-Semitism sent June 3rd to MIT President Sally Kornbluth, MIT Corp Chair, Mark Gorenberg
Dear Dr. Kornbluth and Mr. Gorenberg,
The US House of Representatives is deeply concerned by ongoing and pervasive acts of antisemitic
harassment and intimidation at the Massachusetts Institute of Technology (MIT). Failing to act decisively to ensure a safe learning environment for all students would be a grave dereliction of your responsibilities as President of MIT and Chair of the MIT Corporation.
This Congress will not stand idly by and allow an environment hostile to Jewish students to persist. The House believes that your institution is in violation of Title VI of the Civil Rights Act, and the inability or
unwillingness to rectify this violation through action requires accountability.
Postsecondary education is a unique opportunity for students to learn and have their ideas and beliefs challenged. However, universities receiving hundreds of millions of federal funds annually have denied
students that opportunity and have been hijacked to become venues for the promotion of terrorism, antisemitic harassment and intimidation, unlawful encampments, and in some cases, assaults and riots.
The House of Representatives will not countenance the use of federal funds to indoctrinate students into hateful, antisemitic, anti-American supporters of terrorism. Investigations into campus antisemitism by the Committee on Education and the Workforce and the Committee on Ways and Means have been expanded into a Congress-wide probe across all relevant jurisdictions to address this national crisis. The undersigned Committees will conduct oversight into the use of federal funds at MIT and its learning environment under authorities granted to each Committee.
• The Committee on Education and the Workforce has been investigating your institution since December 7, 2023. The Committee has broad jurisdiction over postsecondary education, including its compliance with Title VI of the Civil Rights Act, campus safety concerns over disruptions to the learning environment, and the awarding of federal student aid under the Higher Education Act.
• The Committee on Oversight and Accountability is investigating the sources of funding and other support flowing to groups espousing pro-Hamas propaganda and engaged in antisemitic harassment and intimidation of students. The Committee on Oversight and Accountability is the principal oversight committee of the US House of Representatives and has broad authority to investigate “any matter” at “any time” under House Rule X.
• The Committee on Ways and Means has been investigating several universities since November 15, 2023, when the Committee held a hearing entitled From Ivory Towers to Dark Corners: Investigating the Nexus Between Antisemitism, Tax-Exempt Universities, and Terror Financing. The Committee followed the hearing with letters to those institutions on January 10, 202
Normal Labour/ Stages of Labour/ Mechanism of LabourWasim Ak
Normal labor is also termed spontaneous labor, defined as the natural physiological process through which the fetus, placenta, and membranes are expelled from the uterus through the birth canal at term (37 to 42 weeks
Synthetic Fiber Construction in lab .pptxPavel ( NSTU)
Synthetic fiber production is a fascinating and complex field that blends chemistry, engineering, and environmental science. By understanding these aspects, students can gain a comprehensive view of synthetic fiber production, its impact on society and the environment, and the potential for future innovations. Synthetic fibers play a crucial role in modern society, impacting various aspects of daily life, industry, and the environment. ynthetic fibers are integral to modern life, offering a range of benefits from cost-effectiveness and versatility to innovative applications and performance characteristics. While they pose environmental challenges, ongoing research and development aim to create more sustainable and eco-friendly alternatives. Understanding the importance of synthetic fibers helps in appreciating their role in the economy, industry, and daily life, while also emphasizing the need for sustainable practices and innovation.
How to Make a Field invisible in Odoo 17Celine George
It is possible to hide or invisible some fields in odoo. Commonly using “invisible” attribute in the field definition to invisible the fields. This slide will show how to make a field invisible in odoo 17.
Francesca Gottschalk - How can education support child empowerment.pptxEduSkills OECD
Francesca Gottschalk from the OECD’s Centre for Educational Research and Innovation presents at the Ask an Expert Webinar: How can education support child empowerment?
Unit 8 - Information and Communication Technology (Paper I).pdfThiyagu K
This slides describes the basic concepts of ICT, basics of Email, Emerging Technology and Digital Initiatives in Education. This presentations aligns with the UGC Paper I syllabus.
A Computational Study Of Traffic Assignment Algorithms
1. Australasian Transport Research Forum 2013 Proceedings
2 - 4 October 2013, Brisbane, Australia
Publication website: http://www.patrec.org/atrf.aspx
1
A Computational Study of Traffic Assignment Algorithms
Olga Perederieieva1
, Matthias Ehrgott2
, Judith Y. T. Wang3
, Andrea Raith1
1
The University of Auckland, Auckland, New Zealand
2
Lancaster University, Lancaster, UK
3
The University of Leeds, Leeds, UK
Email for correspondence: o.perederieieva@auckland.ac.nz
Abstract
Traffic congestion is an issue in most cities worldwide. One way to model and analyse the
effect of congestion and other factors on route choice behaviour and to predict the impact of
traffic management projects and transport policies is traffic assignment (TA). The most
commonly used TA model is known as user equilibrium (UE), which is based on the
assumption that all drivers want to minimise their travel time or generalised cost. As a result,
an equilibrium is achieved when no one has an incentive to switch to another route.
Although the conventional mathematical model of TA belongs to the convex optimisation
domain and, hence, is relatively easy to solve, efficient algorithms are required in order to be
able to solve TA in a reasonable amount of time for realistic transport networks. This
motivates researchers to propose numerous methods and algorithms to solve this problem in
the literature. However, there is no comprehensive empirical study that compares the
performance of different approaches on benchmark instances. In this study, our objective is
to fill this gap.
We provide a literature review of the most promising methods. We classify algorithms
according to the way the solution is represented, namely, link-based (solution is represented
by link flows), path-based (solution is represented by path flows) and origin-based (solution is
represented by link flows corresponding to each origin), and implement the most
representative algorithms in each group. We perform numerical tests on benchmark
instances of various sizes, compare the algorithms and analyse the impact of their main
components on their running time. We also study the convergence behaviour of the methods
with respect to different levels of solution accuracy.
1. Introduction and Motivation
Due to the fast development of cities and road networks the role of transportation planning
grows every day since it provides tools for developing effective transportation spending and
policies. As presented in Ortúzar and Willumsen (2001), transportation planning allows to:
analyse the current usage of a road network; predict the impact of potential projects and
policies; control traffic (level of congestion, emissions, toll revenue etc). In order to achieve
these goals, a model that can realistically describe travel decisions made by drivers is
required. However, in order to create such a model one should know how many people are
travelling, from where to where they are travelling, which travel mode they choose (car, bus,
bicycle etc) and which routes they prefer. Since it is very difficult to predict how a particular
individual will travel, the conventional approach to tackle this difficulty is to make some
assumptions on how people choose routes and to find a flow pattern satisfying these
assumptions. The most well-known such assumptions are the ones following Wardrop's first
principle that is also called user equilibrium condition, (Wardrop, 1952): The journey times on
all the routes actually used are equal, and less than those which would be experienced by a
single vehicle on any unused route.
This principle models the behaviour of travellers by assuming that all drivers are selfish and
they tend to choose the fastest routes going from their origin to their destination. As a result
an equilibrium state is achieved, when no one has an incentive to switch to another route.
2. 2
This principle allows different mathematical formulations of the problem based on different
additional assumptions. The classical model that is commonly used in practice is a static
deterministic traffic assignment (TA) and is the subject of this paper.
This classical model was developed in the 1950s and since then various algorithms were
proposed to solve it. The wide research interest in this problem has several reasons. First,
TA is a challenging problem that arises in different practical applications. Second, due to the
fast development of hardware and software, the existing transportation models continue to
grow. For example, in 2006 the ART model (Auckland Regional Transport Model) included
202 zones, whereas in 2008 it contained 512 zones, (Davies et al., 2009). Third, high
accuracy of the solution is required for the select link analysis1
and for consistent comparison
between design scenarios, as presented in Slavin et al. (2010) and Gentile (2009).
Therefore, we can conclude that there is a growing need for efficient algorithms able to solve
TA problems of realistic size with high accuracy.
Regardless the number of proposed algorithms in the literature, there is no comprehensive
survey study comparing them. According to our knowledge, only the paper of Inoue and
Maruyama (2012) analyses many different methods under the same computational
environment. The authors implemented 11 algorithms for traffic assignment. However, all
implementation details are omitted in the paper. It is not clear if the authors used the same
framework and how carefully they followed the descriptions of the algorithms available in the
literature.
The aim of our research is to analyse and compare the most promising approaches for
solving TA in the context of a framework that can be shared by different algorithms. We want
to compare them without considering any special implementation details that may have been
used in commercial or research software. Instead we focus on the use of common code
wherever possible. This will allow testing general ideas of the methods without giving an
advantage to any of them. We do not claim the best possible implementations, but rather
concentrate on the code that can be shared by as many algorithms as possible and compare
the programming blocks that are different for each particular method.
Another motivation of this study is to identify the advantages and disadvantages of different
groups of algorithms (the classification is presented in Section 3). Early on only link-based
approaches were available in software because of memory limitations. In recent years,
however, other types of algorithms have been implemented in many commercial software
packages used by practitioners. One of the main advantages of new algorithms is that they
can provide the path choice information which is necessary to evaluate effects of schemes
such as congestion pricing without re-running the algorithm, (Florian et al., 2009). Therefore,
it becomes important to analyse them and to compare them with classical link-based
approaches.
The rest of the paper is organised as follows. Section 2 states the deterministic static traffic
assignment problem. Section 3 is devoted to the literature overview and the description of
our choice of the algorithms. In Section 4 various methods for solving traffic assignment are
described. Section 5 discusses the computational study and comparison of the algorithms.
Finally, Section 6 presents conclusion and future work.
2. Problem Formulation
In order to formulate the TA problem the following input data is provided:
1
Select link analysis provides routeing analysis (i.e. a picture of where traffic is coming from and going
to) for assigned vehicles at selected links (and combination of links) throughout the modelled network,
see http://www.transportscotland.gov.uk/analysis/LATIS/data/Model-data/Operational-Analysis.
3. A Computational Study of Traffic Assignment Algorithms
3
A transportation network is defined as a directed graph , where is a set of
nodes and is a set of links;
is the demand of origin-destination (O-D) pair ∈ (demand represents how
many vehicles are travelling from an origin to a destination), is the set of all O-D
pairs;
is a function that describes travel time on link , it depends on link flows
, ,.. . , | | , i.e. the number of vehicles per time unit on each link.
Let F F ,F ,.. . , F| | denote a vector of path flows, where K is the set of all simple paths of
graph G A, N . They are related to link flows by the expression f ∑ ∑ δ
∈
∈ F , where
δ F equals one if link a belongs to path k, and zero otherwise; K ⊆ K is the set of paths
between O-D pair p. Let C F denote the travel time on path k, it is also called path cost
function.
The conventional model of the traffic assignment problem is based on two important
assumptions that allow to formulate and solve it as a mathematical program. These
assumptions are stated as follows:
Additivity of path cost functions: travel time of each path is the sum of travel times of
links belonging to this path, i.e. ∑ ∈ ;
Separability of link cost functions: travel time of each link depends only on flow on this
link, i.e. .
If these assumptions are satisfied, solving the following optimisation problem results in the
link flows satisfying the user equilibrium condition, (Sheffi, 1985):
min c
∈
x dx
F
∈
D , ∀p ∈ Z,
F ⩾ 0, ∀k ∈ K , ∀p ∈ Z,
f δ
∈
∈
F , ∀a ∈ A.
(2.1)
In order to ensure the existence of a solution of TA all path cost functions C F must be
positive and continuous, and to ensure the uniqueness of the solution of TA in terms of link
flows f all path cost functions C F must be strictly monotone (Florian and Hearn, 1995). In
the following it is assumed that these requirements are satisfied.
3. Literature Overview
One of the possible ways to classify traffic assignment algorithms is according to how the
solution is represented: link-based (solution variables are link flows), path-based (solution
variables are path flows) and origin-based (solution variables are link flows coming from a
particular origin), see (Zhou and Martimo, 2010).
Historically the first algorithms developed for solving the traffic assignment problem were
link-based. The most well-known such algorithm is Frank-Wolfe (FW), (Frank and Wolfe,
1956). Due to its simplicity and low memory requirements it is used even now and is
implemented in different commercial software packages. However, this algorithm and other
link-based methods are known to tail badly in the vicinity of the optimum and usually cannot
be used to achieve highly precise solutions as is demonstrated by numerous numerical
studies, see Lee et al. (2002), Florian et al. (2009) and Inoue and Maruyama (2012).
4. 4
Many improvements of FW were proposed in the literature. As explained in Zhou and
Martimo (2010), some of them try to improve the FW search direction (for example, see
Fukushima (1984), Florian et al. (1987), Mitradjieva and Lindberg (2012)) or step size (see
Powell and Sheffi (1982), Weintraub et al. (1985)). Zhou and Martimo (2010) also categorise
restricted simplicial decomposition (RSD), (Hearn et al., 1985) and nonlinear simplicial
decomposition (NSD), (Larsson et al., 1997) as link-based methods. These algorithms apply
a more complicated structure than FW, however they use FW as a special case. In both of
them the link flow solution is represented by a linear combination of extreme points. The
algorithms of this type usually have two main routines: generating new extreme points and
optimising the so-called restricted master problem with respect to previously generated
extreme points.
The first path-based algorithm was proposed in Dafermos and Sparrow (1969). It is called
path equilibration (PE). Its main idea is to shift flow from the path with maximum cost to the
path with minimum cost. However, in the years after the PE algorithm was published, path-
based methods were considered impractical because of memory limitations. As a result, the
development in this field started again only in recent years. Larsson and Patriksson (1991)
proposed disaggregated simplicial decomposition (DSD) which is similar to RSD in the way
that it also uses a linear combination of extreme points to represent the solution. However,
the extreme points are in the space of path flows instead of link flows. In Jayakrishnan et al.
(1994) the gradient projection (GP) method was proposed and further studied in Chen and
Jayakrishnan (1998). It is similar to PE, but O-D flow is moved from several non-shortest
paths to the shortest path. Another path-based algorithm was proposed in Florian et al.
(2009). It is called projected gradient (PG) and is based on the idea of moving flow from the
set of paths with cost greater than the average path cost to the set of paths with cost less
than the average path cost. Another similar approach called improved social pressure (ISP)
algorithm was developed in Kumar and Peeta (2011). This method also shifts flow from the
set of costlier paths to the set of cheaper paths, but a more complicated strategy of flow
distribution is applied.
The origin-based algorithms represent a more recent development in this field. Their main
idea is to decompose the problem into a sequence of subproblems that operate on acyclic
subnetworks of the original transportation network (Nie, 2010). In general, for this group of
algorithms, the flow shifts are restricted to subproblems and are usually similar to the ideas
presented earlier for path- and link-based approaches. The first algorithm of this type applied
to the traffic assignment problem was proposed by Bar-Gera (2002). It is called origin-based
algorithm (OBA). Nie (2011) presented some corrections to OBA, and Nie (2010) compared
different origin-based algorithms (corrected OBA (COBA) and modified OBA (MOBA)) based
on OBA. Other developments in this field include the following methods: algorithm B (B)
proposed by Dial (2006), some modifications of it (iB), see Inoue and Maruyama (2012) and
Tianran et al. (2010), linear user cost equilibrium (LUCE) developed by Gentile (2009) and
traffic assignment by pairs of alternative segments (TAPAS) implemented by Bar-Gera
(2010).
In order to select algorithms for implementation we analyse different empirical studies from
the literature. During such analysis it is important to pay attention to how the algorithms were
compared (re-implemented by the authors of the study or using existing software), what
instances were used and how precise the obtained solutions were.
In the majority of the studies relative gap is used as a convergence measure. It is calculated
as follows:
RGAP 1
∑ D
∈ C
∑ f
∈ c
(3.1)
5. A Computational Study of Traffic Assignment Algorithms
5
where C min ∈ C is the shortest path of O-D pair p. We divide the existing numerical
studies into two main groups (low and high precision) corresponding to the accuracy of the
solution. The algorithms from the low precision group are stopped when the relative gap is in
the interval 10 , 10 , and for the high precision group the interval is 10 , 10 . Table
1 presents the first group and Table 2 the second one. In each table one algorithm from each
study is highlighted in bold, which means that it showed the best performance on the majority
of the tested instances. All other algorithms in the same study are written in the order of
decreasing performance. If in a particular study the existing executable of the algorithm or
commercial software was used, it is highlighted in grey. If comparison of the algorithms was
made indirectly, i.e. based on the running times reported in other papers, it is highlighted in
light grey. We also try to align the algorithms that were compared in different studies.
However, since sometimes contradictory conclusions are made in different papers, this
alignment is very approximate. The last line of each table summarises the algorithms that
seem to be the most promising, namely BFW (bi-conjugate Frank-Wolfe (Mitradjieva and
Lindberg, 2012)), GP, PG, ISP, B, LUCE, TAPAS.
Therefore, we choose to implement the following algorithms:
Link-based: FW (FW is reported to have the worst performance. However, it is the
basis for CFW (conjugate Frank-Wolfe (Mitradjieva and Lindberg, 2012)) and BFW,
that is why it is also considered in our study), CFW, BFW;
Path-based: PE (this algorithm was not compared in the studies presented above.
However, we choose to implement it as well due to its simplicity and ideas similar to
other path-based approaches), GP, PG and ISP;
Origin-based: B.
The most recent algorithms that also show very promising performance, namely LUCE and
TAPAS, are subject of our future research.
Table 1: Summary of existing empirical studies. Low precision: ∈ ,
Mitradjieva
and
Lindberg
(2012)
Zhou
and
Martimo
(2010)
Slavin et
al.
(2006)
Dial
(2006)
Florian
et al
(2009)
Gentile
(2009)
Tianran
et al.
(2010)
Kumar
and
Peeta
(2011)
LUCE ISP
PG PG
SP (older
version of
ISP)
B B B iB, B
OBA OBA OBA OBA
FW FW FW FW FW
GP GP OBA
BFW BFW
CFW CFW
FW FW
OBA OBA
DSD
Best: BFW, GP, PG, ISP, B, LUCE
6. 6
Table 2: Summary of existing empirical studies. High precision: ∈ ,
Bar-Gera (2002) Nie (2010) Bar-Gera (2010) Inoue and Maruyama (2012)
TAPAS TAPAS
B B, iB
MOBA
COBA
OBA OBA OBA OBA, DSD, MOBA, LUCE, ASD
FW FW FW
Best: B, TAPAS
4. Algorithms
One of the reasons for the development of various traffic assignment algorithms is a specific
problem structure that can be exploited by solution methods in many different ways
(Patriksson, 1994).
The constraints of the TA problem represent a polyhedral set and the objective function is
convex. As a result feasible direction methods can be applied as a solution technique
(Andréasson et al., 2007). The algorithms described later in this section belong to this type of
nonlinear optimisation methods. The main idea behind this type of methods is as follows:
starting from a feasible solution, a feasible direction of descent is calculated and the solution
is moved along this direction in such a way that the strict decrease of the objective function is
guaranteed.
Another property of traffic assignment that can be exploited is the O-D pair separability: the
flow conservation constraints of one O-D pair do not affect those of any other pair. This
allows applying decomposition algorithms, that are based on the idea of decomposing the
problem into smaller sub-problems that can be solved one after another or in parallel, see
Patriksson (1994). This section discusses different decomposition approaches and
algorithms. In particular we focus on the algorithms that were chosen for our numerical study,
see Section 3.
4.1. Link-based Algorithms
This Section presents Frank-Wolfe and two of its modifications: conjugate and bi-conjugate
Frank-Wolfe methods. These algorithms can be described using the framework presented in
Figure 1. As can be seen from the framework the methods differ only in the way the direction
of descent is defined.
Each algorithm starts with an initial feasible link flow pattern. It is usually generated using all-
or-nothing (AON) assignment: each link flow is initialised with zero and corresponding link
travel times are calculated, then shortest paths for each O-D pair are found and
corresponding demand is assigned to each shortest path. All these path flows are then
projected on all links resulting in the initial feasible link flows.
FW further exploits the AON procedure. In order to generate a feasible direction of descent it
performs AON assignment at each iteration and gets corresponding link flows y , ∀a ∈ A.
Since the AON link flow solution is feasible, given a current link flow solution f , ∀a ∈ A, vector
7. A Computational Study of Traffic Assignment Algorithms
7
d y f , ∀a ∈ A is a feasible direction. Moreover, this direction is also a direction of
descent (Sheffi, 1985).
Figure 1: Framework for link-based algorithms
CFW and BFW algorithms use information about previously generated directions of descent
in order to find a new one. In particular, the new search direction is constructed via mutually
conjugate directions with respect to the Hessian of the objective: CFW takes into account the
usual FW direction and the direction from the previous iteration, whereas BFW in addition
considers one more direction from iteration i 2 where i is the current iteration. Both
algorithms operate on auxiliary variables s , ∀a ∈ A called points of sight that store
information about previously generated directions of descent. For details see Mitradjieva and
Lindberg (2012).
After finding a direction of descent each algorithm proceeds to step size calculation which
represents how far away the current solution must be moved in the direction of descent. The
step size can be found by solving exactly or approximately the following optimisation problem
(Florian and Hearn, 1995):
min ⩽ ⩽ c
∈
ω dω (4.1)
There are various methods available for solving the one-dimensional optimisation problem
(4.1), see for example Andréasson et al. (2007). In this study we applied quadratic
interpolation, i.e. the objective function along the direction of descent is approximated by a
quadratic function and minimised using the analytic solution:
λ
∑ c
∈ f d
∑
∂c f
∂f
d
∈
.
(4.2)
3.4. Path-based Algorithms
Path-based methods exploit the O-D pair separability of the TA problem. At each iteration,
the flows are moved only within one O-D pair and path flows of the other O-D pairs are fixed.
In order to achieve this, paths must be stored. Let K , ∀p ∈ Z denote the corresponding sets
of paths. A general framework of this group of algorithms is presented in Figure 2.
In order to prevent storing all possible simple paths for each O-D pair a column generation
approach is usually applied, which consists in generating new paths when needed
(Patriksson, 1994). In particular, it is performed as follows: for a given O-D pair p find the
shortest path and add it to K if it is shorter than the current shortest path contained in this
set. This step corresponds to “Improve path set “ of the framework. Also, in order to keep
only promising paths in K , the paths that do not carry flow anymore are removed.
8. 8
Figure 2: Framework for path-based algorithms
As in the case of link-based methods, initialisation is performed by AON assignment. In
addition to a link flow solution each set K , ∀p ∈ Z of paths must also be initialised. This can
be done by adding the shortest path corresponding to O-D pair p to K .
All algorithms of this group differ in how path set K is equilibrated. Equilibration of the set
aims at equalising some or all of the path costs of set K . The remainder of this section
discusses different algorithms belonging to this group.
The PE algorithm equalises the costs of the current longest path l ∈ K with positive flow and
the cost of the current shortest path s ∈ K . This is equivalent to solving a non-linear
equation. We apply Newton's method in order to solve it, namely we calculate the Newton
step ΔF
∑ ∈ ,
, where A , is the set of links that belong to path l and path s without links
common to both these paths, and update path flows in the following way:
F F min F , ΔF ,
F F min F , ΔF .
(4.3)
Expression min F , ΔF is necessary, because the calculated flow shift ΔF might be infeasible
causing F to become negative.
The GP algorithm considers several paths at each iteration. In particular, it moves flow to the
current shortest path s ∈ K from all other paths of set K . First, an appropriate flow shift is
calculated:
ΔF
C C
∑
∂c
∂f
∈ ,
, ∀k ∈ K , k s.
(4.4)
Second, a new solution is projected onto the feasible set as follows:
F F min F , αΔF , ∀k ∈ K , k s,
F D F
∈ ,
, (4.5)
where α is a predefined constant that must be small enough in order to guarantee
convergence of the algorithm (Jayakrishnan et al., 1994). In our study it was set to 0.25
because this value allows all tested instances to converge (Jayakrishnan et al. (1994)
recommend to set it to 1).
9. A Computational Study of Traffic Assignment Algorithms
9
The main idea of the PG algorithm is to move flow from the paths that have cost greater than
the current average path cost to the paths that have cost less than the average value. It is
equivalent to defining the following direction of descent d:
d C C , ∀k ∈ K (4.6)
where C
∑ ∈
is the average cost of the paths of O-D pair p. In order to find the
appropriate amount of flow to move line search is applied along direction d. As a result, step
size λ is calculated. We used quadratic interpolation for this purpose, see Section 4.1. The
path flows are updated accordingly: F F λd , ∀k ∈ K .
The ISP algorithm is based on the idea of “social pressure” which is defined as the difference
between a path cost and the cost of the shortest path. All the paths are divided into two
groups P ⊂ K (paths with the cost less than or equal to some value π) andP ⊂ K (paths
with the cost greater than π) such that P ∪ P K . π is determined at each iteration and is
defined as follows: π C δ C C , where C and C are the costs of current shortest and
longest paths, δ is a predefined constant that was set to 0.15 as suggested in the paper of
Kumar and Peeta (2011). Flow is shifted from the paths belonging to set P to the paths
belonging to set P . This is also equivalent to defining a direction of descent d and moving
the solution along this direction. Direction d is as follows:
d C C , ∀k ∈ P ,
d
∑ Δ
∈ F
s F ∑
1
s F
∈
, ∀l ∈ P , (4.7)
where s F is the first derivative of the cost function of the path m with respect to path
flow: s F ∑ ∈ . The step size λ is calculated as presented in Section 4.1 and
path flows are updated as follows: F F λd , ∀k ∈ K .
4.3 Origin-based Algorithms
Origin-based algorithms also exploit the O-D pair separability of the TA problem. But instead
of decomposing the problem into sub-problems corresponding to each O-D pair, the
algorithms of this type decompose the problem into sub-problems corresponding to each
origin. As a result, at each iteration flows are moved within one origin using a special data
structure called bush. A bush is a directed sub-network of the original network G A, N rooted
at a given origin o such that it is (Nie, 2010):
Connected, i.e. using only links in the bush it is possible to reach every node that was
reachable in the original network;
Acyclic, i.e. the bush does not contain directed cycles.
We denote a bush by B N, A ⊂ G A, N , where A ⊂ A. Let O ⊆ N denote the set of origins.
Then, at each iteration only one bush B N, A is considered. The current link-based solution
f is the sum of |O| link flows of B N, A , ∀o ∈ O, i.e. f ∑ f
∈ , ∀a ∈ A, where f denotes
the flow on link a of the bush B N, A (Dial, 2006).
10. 10
Figure 3: Framework for origin-based algorithms
The motivation to decompose the traffic assignment problem by origins consists in the
acyclicity of user equilibrium: for the single-origin formulation of the traffic assignment
problem, links that have positive flow at user equilibrium never form a directed cycle (Nie,
2010). This property gives several interesting advantages from a practical point of view since
different operations such as shortest path calculations can be performed more efficiently on
acyclic structures (such as bushes) than on general networks (Nie, 2010).
All algorithms that decompose the TA problem by origins share a general framework
presented in Figure 3.
Each bush B N, A is usually initialised with a shortest path tree rooted at origin o and the
link flows are initialised by AON assignment.
As in the case of path-based methods, bushes are constructed iteratively by adding
promising links and removing unused ones. The removal of links is performed in the following
way: links that carry zero flow are dropped if the connectivity of the bush is retained. The
addition of new links must be performed with care in such a way that directed cycles are not
created. The reader is referred to Nie (2010) for a detailed explanation of this algorithmic
step.
In our study we consider only one origin-based algorithm called algorithm B, see Dial (2006).
Its main idea is similar to PE (see Section 4.2), namely the flow is shifted from the longest
used path to the shortest path, but within a given bush. We implemented the same approach
as in the case of the PE algorithm: the Newton step is applied to equalise the path costs. It
must be noticed that the explicit storage of paths is not required anymore, because the
calculation of the longest path can be done in linear time on the bush due to its acyclicity. In
the case of path-based algorithms the paths must be stored because the longest path
calculation is NP-complete on general graphs (Patriksson, 1994).
5. Computational Study
5.1 Environment and Problem Instances
This section summarises our empirical study. All algorithms were implemented in the C++
programming language and compiled using g++ 4.7.3 (Ubuntu/Linaro 4.7.3-1ubuntu1). All
runs were performed under the following environment:
11. A Computational Study of Traffic Assignment Algorithms
11
OS: Ubuntu Release 13.04 64-bit;
CPU: Intel Core i5-2500 CPU, 4 Core, 3.30GHz;
RAM: 7.7 GB.
The tests were performed on the instances available at the web-site:
http://www.bgu.ac.il/~bargera/tntp/. The main characteristics of these instances are
presented in Table 3.
Table 3: Problem instances
Instance name Nodes Links Zones O-D pairs
Sioux-Falls 24 76 24 528
Anaheim 416 914 38 1406
Barcelona 1020 2522 110 7922
Winnipeg 1052 2836 147 4344
Chicago Sketch 933 2950 387 93135
We use the relative gap RGAP (see Section 3) as a convergence criterion for two reasons: it
is a common measure of convergence, see Slavin et al. (2006), Florian et al. (2009), Dial
(2006), Nie (2010), and it can be calculated for all tested algorithms. The required accuracy
of the solution was set to ϵ 10 , i.e. the algorithms were stopped after the relative gap
was less than ϵ.
We have also performed one more numerical test on the Chicago Regional network that is
much larger than the instances presented in Table 3. It has 1791 nodes, 39018 links, 1790
zones and 2296227 O-D pairs. For this test we set a time limit of 8 hours for every algorithm.
The results are discussed in the following sections.
5.2 Implementation
All algorithms were implemented in terms of the frameworks presented in Section 4. This
ensures the usage of common code wherever possible.
Each TA method requires solving the shortest path problem multiple times. Link-based
approaches need a single-source shortest path algorithm for general graphs. The same
algorithm is also required for the relative gap calculation. For this purpose we implemented
the label correcting algorithm presented in Sheffi (1985). Path-based approaches, on the
other hand, require a point-to-point shortest path method since at each iteration only one O-
D pair is considered. For this case we used the A* algorithm2
, see Goldberg et al. (2006). For
origin-based algorithm B we implemented the conventional shortest path algorithm for
directed acyclic graphs that uses topological order, see Dasgupta et al. (2006). It should also
be noticed that each bush is constructed iteratively. This means that topological sorting can
be performed dynamically (Pearce and Kelly, 2006). However, our current implementation
does not exploit this fact and the topological order is computed from scratch each time the
bush topology changes.
In Florian et al. (2009) it is mentioned that when the solution is close to the optimal one,
many O-D pairs are equilibrated and there is no need to change their path flows and improve
path sets. Our implementation does not take into account this empirical fact. Consequently,
some algorithmic steps are performed even in the case when this is not necessary. For
2
An implementation of the A* algorithm was kindly provided by Boshen Chen.
12. 12
example, the improvement of path set K which involves calculation of the shortest path is
performed at each iteration regardless how close to equilibrium O-D pair p is.
Also our current implementation is not optimised with respect to memory usage. For all
algorithms we used extended floating point precision (C++ long double type).
5.3 Results
Our first numerical test was performed on the small and medium size instances presented in
Table 3. All link-based methods were not able to reach the required accuracy of relative gap
in a reasonable amount of time (execution time of all algorithms was limited to 8 hours). The
PG algorithm showed unstable numerical behaviour consisting in the loss of flow at each
iteration. Basically, because of rounding errors in the computation of the average path cost,
the amount of flow lost at each iteration prevents this algorithm from converging when the
required precision is approximately less than 10 (this value depends on the problem
instance).
Figure 4 presents running times of the algorithms that were able to reach the required
precision. According to these results, algorithm B is usually significantly faster than other
methods. PE shows the best performance among path-based algorithms followed by ISP and
GP. GP is usually slower because its performance depends on parameter α. In our numerical
tests we didn't adjust this parameter to improve performance of the algorithm since it is
instance-dependent. We think that ISP is slower than PE mainly because it requires
additional computations at each step, see Section 4.2.
Figures 5 and 7 show convergence of the algorithms on the Barcelona and Chicago Sketch
instances, respectively, whereas Figures 6 and 8 show early stages of the same
convergence. After analysing these results, we can conclude that link-based methods
converge fast on early iterations, but start tailing in the vicinity of the optimum. Usually, if the
required level of precision is 10 of relative gap, link-based methods represent a reasonable
alternative to other approaches. Among the algorithms of this type BFW demonstrated the
best performance on the majority of the tested instances. If the required precision is stricter,
then algorithm B is a good choice in terms of both memory usage and running time
compared to path-based approaches.
Different convergence patterns of different groups of algorithms occur in Figures 5 and 7. In
particular, all link-based methods tail in the vicinity of optimum. As mentioned in Patriksson
(1994), Jayakrishnan et al. (1994) and Mitradjieva and Lindberg (2012), this happens
because the search direction is perpendicular to the gradient of the objective function when
the algorithm approaches the optimal solution. This does not happen for path- and origin-
based algorithms. In Figure 5 the relative gap of algorithm B and PE drops by a factor of 10
after reaching a value of 10 . This behaviour is not common (for example, see Figure 7).
One possible explanation of this phenomenon is the fact that the relative gap is measured
after all O-D pairs are considered and not after each flow move. Therefore, for this particular
instance, the relative gap might reduce by a factor of 10 after flows are shifted within all O-D
pairs.
Our second numerical test was performed on a large instance of the Chicago Regional
network. The required precision was set to 10 of relative gap and time of execution was
limited to 8 hours. All path-based algorithms were not able to reach the required precision
because of the lack of memory. One obvious reason is absence of memory usage
optimisation of the current implementation. The convergence of the algorithms is presented
in Figure 9, with a few first iterations of path-based algorithms before they run out of memory
(the PG algorithm is not presented because it ran out of memory after the first iteration). As
can be seen from the results, the fastest algorithm was BFW. Algorithm B starts dominating
FW and CFW when the relative gap is less than 10 . However, on this level of solution
accuracy it is still slower than BFW.
13. A Computational Study of Traffic Assignment Algorithms
13
5.4 Summary
Each type of algorithm presented above has its advantages and disadvantages that must be
taken into account when solving the TA problem. Depending on the particular requirements
on the solution, a problem instance and available resources, preference must be given to a
particular algorithm. Table 4 presents some potential trade-offs of different approaches.
6. Conclusion and Future Work
This work starts with a literature overview of existing traffic assignment algorithms, based on
which we selected and implemented the most promising ones. The main aim of our
numerical study is to compare the algorithms under the same computational environment
and in terms of a framework that insures that common code is used wherever possible. We
implemented and analysed several algorithms belonging to link-, path- and origin-based
methods, namely FW, CFW, BFW, PE, GP, PG, ISP and B.
Based on the performed computational study we can conclude that link-based algorithms
represent a good alternative to other methods when the required level of precision is low.
Path-based algorithms must be implemented with care in terms of memory usage, especially
if the problem instance is large. Origin-based algorithm B is capable to achieve a high level
of precision with moderate memory consumption and is preferable when high precision of the
solution is needed. Numerical stability of the algorithms is also an important issue especially
if high accuracy of the solution is necessary and must be taken into account.
The future development of this study consists in implementing the most recent algorithms
such as LUCE and TAPAS and in improving the current code by reducing memory usage of
path-based algorithms, adding on-line topological sorting and other TA specific features that
can be exploited by different methods. Also we plan to study the performance of algorithms
with respect to the representative operation counts in order to identify and analyse the
asymptotic bottleneck operations.
Figure 4: Running time of algorithms
14. 14
Table 4: Trade-offs of different TA algorithms
Factor Link-based Path-base Origin-based
Memory
requirements
Low: | |
High: potentially exponential
(number of paths in graph is
exponential)
Moderate: | || |
Solution
precision
Low: usually cannot
achieve high precision in
reasonable amount of time
High High
Shortest path
algorithm
Single-source shortest path
on general graph
Point-to-point shortest path
on general graph
Single-source shortest
path on directed
acyclic graph
Flow moves at
each iteration
Flows are moved within
entire network
Flows are moved within path
set
Flows are moved
within bush
Convergence
behaviour
Quickly converges in the
beginning, tails near vicinity
of optimum
Do not tail in the vicinity of optimum
Figure 5: Convergence of algorithms. Barcelona instance
15. A Computational Study of Traffic Assignment Algorithms
15
Figure 6: Convergence of algorithms on early stages. Barcelona instance
Figure 7: Convergence of algorithms. Chicago Sketch instance
16. 16
Figure 8: Convergence of algorithms on early stages. Chicago Sketch instance
Figure 9: Convergence of algorithms. Chicago Regional instance
17. A Computational Study of Traffic Assignment Algorithms
17
References
N. Andréasson, A. Evgrafov, and M. Patriksson. An Introduction to Continuous Optimization.
Studentlitteratur, 2007.
H. Bar-Gera. Origin-based algorithm for the traffic assignment problem. Transportation
Science, 36(4):398–417, 2002.
H. Bar-Gera. Traffic assignment by paired alternative segments. Transportation Research
Part B: Methodological, 44(8-9):1022 – 1046, 2010.
A. Chen and R. Jayakrishnan. A path-based gradient projection algorithm: effects of
equilibration with a restricted path set under two flow update policies. In Transportation
Research Board Annual Meeting, 1998.
S. C. Dafermos and F.T. Sparrow. The traffic assignment problem for a general network.
Journal of research of the National Bureau of Standards. B, Mathematical sciences., 73B:91–
118, 1969.
S. Dasgupta, C.H. Papadimitriou, and U.V. Vazirani. Algorithms. McGraw-Hill Education
(India) Pvt Limited, 2006.
J. Davies, J. Valero, and D. Young. The Auckland transport models project: Overview and
use to date. In Australasian Transport Research Forum 2009 Proceedings, 2009.
R. Dial. A path-based user-equilibrium traffic assignment algorithm that obviates path storage
and enumeration. Transportation Research Part B: Methodological, 40(10):917–936, 2006.
M. Florian and D. Hearn. Network equilibrium models and algorithms. In C.L. Monma M.O.
Ball, T.L. Magnanti and G.L. Nemhauser, editors, Network Routing, volume 8 of Handbooks
in Operations Research and Management Science, chapter 6, pages 485 – 550. Elsevier,
1995.
M. Florian, J. Gúelat, and H. Spiess. An efficient implementation of the PARTAN variant of
the linear approximation method for the network equilibrium problem. Networks, 17:319–339,
1987.
M. Florian, I. Constantin, and D. Florian. A new look at projected gradient method for
equilibrium assignment. Transportation Research Record: Journal of the Transportation
Research Board, 2090:10–16, 2009.
M. Frank and P. Wolfe. An algorithm for quadratic programming. Naval Research Logistics
Quarterly, 3(1-2):95–110, 1956. ISSN 1931-9193.
M. Fukushima. A modified frank-wolfe algorithm for solving the traffic assignment problem.
Transportation Research Part B: Methodological, 18(2):169–177, 1984.
G. Gentile. Linear user cost equilibrium: a new algorithm for traffic assignment. Working
paper, April 2009.
A. V. Goldberg, H. Kaplan, and R. F. Werneck. Reach for A: Efficient point-to-point shortest
path algorithms. In Workshop on Algorithm Engineering & Experiments, pages 129–143,
2006.
D. W. Hearn, S. Lawphongpanich, and J. A. Ventura. Finiteness in restricted simplicial
decomposition. Oper. Res. Lett., 4(3):125–130, October 1985. ISSN 0167-6377.
S. Inoue and T. Maruyama. Computational experience on advanced algorithms for user
equilibrium traffic assignment problem and its convergence error. Procedia - Social and
Behavioral Sciences, 43(0):445 – 456, 2012. ISSN 1877-0428. 8th International Conference
on Traffic and Transportation Studies (ICTTS 2012).
18. 18
16R. Jayakrishnan, W. K. Tsai, J. Prashker, and S. Rajadhyaksha. A faster path-based
algorithm for traffic assignment. Transportation Research Record, 1443:75–83, 1994.
A. Kumar and S. Peeta. An improved social pressure algorithm for the static deterministic
user equilibrium traffic assignment problem. Transportation Research Board of the National
Academics, Washington, D.C., 2011.
T. Larsson and M. Patriksson. Simplicial decomposition with disaggregated representation
for the traffic assignment problem. Transportation Science, 26:4–17, 1991.
T. Larsson, M. Patriksson, and C. Rydergren. Applications of simplicial decomposition with
nonlinear column generation to nonlinear network flows. In of Lecture Notes in Economics
and Mathematical Systems, pages 346–373. Springer-Verlag, 1997.
D.-H. Lee, Y. Nie, A. Chen, and Y. C. Leow. Link- and path-based traffic assignment
algorithms: Computational and statistical study. Transportation Research Record: Journal of
the Transportation Research Board, 1783:80–88, 2002.
M. Mitradjieva and P. O. Lindberg. The stiff is moving - conjugate direction Frank-Wolfe
methods with applications to traffic assignment. Transportation Science, Articles in Advance,
INFORMS, pages 1–14, 2012.
Y. Nie. A class of bush-based algorithms for the traffic assignment problem. Transportation
Research, 44:73–89, 2010.
Y. Nie. A note on Bar-Gera’s algorithm for the origin-based traffic assignment problem.
Transportation Science. Under review, 2011.
J. D. Ortúzar and L.G. Willumsen. Modelling Transport. J. Wiley, Chichester New York, 2001.
M. Patriksson. The traffic assignment problem: models and methods. Topics in
transportation. VSP, 1994.
D. J. Pearce and P. H. J. Kelly. A dynamic topological sort algorithm for directed acyclic
graphs. J. Exp. Algorithmics, 11:2006, 2006.
W. B. Powell and Y. Sheffi. The Convergence of Equilibrium Algorithms with Predetermined
Step Sizes. Transportation Science, 16(1), February 1982.
Y. Sheffi. Urban transportation networks: equilibrium analysis with mathematical
programming methods. Prentice-Hall, Englewood Cliffs, N.J., 1985.
H. Slavin, J. Brandon, and A. Rabinowicz. An empirical comparison of alternative user
equilibrium traffic assignment methods. In European Transport Conference, 2006.
H. Slavin, P. Ricotta, J. Brandon, A. Rabinowicz, and S. Sundaram. A new traffic assignment
method for small and medium communities. In Tools of the Trade Conference, September
2010.
Z. Tianran, Y. Chao, and C. Dongdong. Computational study of origin-based user equilibrium
traffic assignment algorithms. In Intelligent Computation Technology and Automation
(ICICTA), 2010 International Conference on, volume 1, pages 1002 –1006, may 2010.
J.G. Wardrop. Some theoretical aspects of road traffic research. Institution of Civil Engineers,
1952.
A. Weintraub, C. Ortiz, and J. Gonzalez. Accelerating convergence of the Frank-Wolfe
algorithm. Transportation Research Part B: Methodological, 19(2):113–122, April 1985.
Z. Zhou and M. Martimo. Computational study of alternative methods for static traffic
equilibrium assignment. In Proc. World Conf. Transport Res. Soc. WCTRS, Lisbon, Portugal,
2010.
View publication stats
View publication stats