Multi agent quality of service routing based on scheme ant colony optimization algorithm
The current Internet needs to support a wide variety of applications with different demands in terms of Quality of Service (QoS) requirements including constraints on throughput, delay, jitter, loss rate. Differentiated Service (DiffServ) has been proposed by the Internet Engineering Task Force (...
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2014
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/76128/ http://psasir.upm.edu.my/id/eprint/76128/1/ITMA%202014%2014%20IR.pdf |
| _version_ | 1848857880708513792 |
|---|---|
| author | Baygi, Maassoumeh Javadi |
| author_facet | Baygi, Maassoumeh Javadi |
| author_sort | Baygi, Maassoumeh Javadi |
| building | UPM Institutional Repository |
| collection | Online Access |
| description | The current Internet needs to support a wide variety of applications with different
demands in terms of Quality of Service (QoS) requirements including constraints on
throughput, delay, jitter, loss rate. Differentiated Service (DiffServ) has been proposed by
the Internet Engineering Task Force (IETF), motivated by the desire to serve different
types of traffic in IP networks. While much of the existing work on DiffServ focus on
scheduling policies, in order to provide QoS for high-quality Internet applications, QoSaware
path selection is important. Known as QoS Routing (QoSR), this multiple
constraints path selection is considered NP-Complete. Some algorithms provide QoS in
term of routing, but they usually focus on improving the performance of the highest
priority traffic, at the expense of the lower priority classes. In order to fully exploit the
network resources, and to meet QoS requirements of various classes concurrently, it is
important to establish class dependent paths. Therefore, an algorithm which can
adaptively assign network resources to meet the QoS requirements of different classes
simultaneously is still needed.
This research presents a per class QoS routing approach based on Ant Colony
Optimization (ACO) called ACR-QoS to provide QoS for different Class of Services
(CoSs). Traffic of different classes can be distributed in accordance with network states
and QoS requirements. The important new aspect of ACR-QoS is the combination of
constraint-based routing and DiffServ architecture, in Swarm Intelligence (SI) structure
where a set of artificial ants is used to determine the optimal path for each class to
construct class-based routing tables. This approach allows easy and efficient design of
service classes, while no changes are needed at lower layers. This study also introduces a
new probe-based procedure for discovery and setup QoS path for real-time traffic. The
proposed scheme has been simulated by OMNET++ and compared with standard AntNet
and two well-known standard QoS routings; Widest Shortest Path (WSP) algorithm and Shortest Widest Path (SWP) algorithm. Experiment results show that the desired service
class differentiation is obtained and ACR-QoS satisfies the QoS requirements of each
CoS without discriminating against the best-effort traffic. In comparison with standard
AntNet, results confirm that the modifications and extensions applied to AntNet can
significantly reduce the average packet delay and jitter by about 11.8% and 46%
respectively in transient regime. It can also improve the throughput in link failure state by
about 25% and decrease the packet loss rate by 27% and 39% before and during link
failure respectively. Moreover, the proposed approach outperforms the WSP and SWP
algorithms, where network efficiency in saturated load gains an 8.39% improvement
while the satisfaction of delay and jitter constraints for time critical applications are also
achieved. In saturated load, efficiency is a very important parameter and a few changes in
it can result in high performance of network to delivery of data. |
| first_indexed | 2025-11-15T12:04:35Z |
| format | Thesis |
| id | upm-76128 |
| institution | Universiti Putra Malaysia |
| institution_category | Local University |
| language | English |
| last_indexed | 2025-11-15T12:04:35Z |
| publishDate | 2014 |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | upm-761282019-11-27T06:16:19Z http://psasir.upm.edu.my/id/eprint/76128/ Multi agent quality of service routing based on scheme ant colony optimization algorithm Baygi, Maassoumeh Javadi The current Internet needs to support a wide variety of applications with different demands in terms of Quality of Service (QoS) requirements including constraints on throughput, delay, jitter, loss rate. Differentiated Service (DiffServ) has been proposed by the Internet Engineering Task Force (IETF), motivated by the desire to serve different types of traffic in IP networks. While much of the existing work on DiffServ focus on scheduling policies, in order to provide QoS for high-quality Internet applications, QoSaware path selection is important. Known as QoS Routing (QoSR), this multiple constraints path selection is considered NP-Complete. Some algorithms provide QoS in term of routing, but they usually focus on improving the performance of the highest priority traffic, at the expense of the lower priority classes. In order to fully exploit the network resources, and to meet QoS requirements of various classes concurrently, it is important to establish class dependent paths. Therefore, an algorithm which can adaptively assign network resources to meet the QoS requirements of different classes simultaneously is still needed. This research presents a per class QoS routing approach based on Ant Colony Optimization (ACO) called ACR-QoS to provide QoS for different Class of Services (CoSs). Traffic of different classes can be distributed in accordance with network states and QoS requirements. The important new aspect of ACR-QoS is the combination of constraint-based routing and DiffServ architecture, in Swarm Intelligence (SI) structure where a set of artificial ants is used to determine the optimal path for each class to construct class-based routing tables. This approach allows easy and efficient design of service classes, while no changes are needed at lower layers. This study also introduces a new probe-based procedure for discovery and setup QoS path for real-time traffic. The proposed scheme has been simulated by OMNET++ and compared with standard AntNet and two well-known standard QoS routings; Widest Shortest Path (WSP) algorithm and Shortest Widest Path (SWP) algorithm. Experiment results show that the desired service class differentiation is obtained and ACR-QoS satisfies the QoS requirements of each CoS without discriminating against the best-effort traffic. In comparison with standard AntNet, results confirm that the modifications and extensions applied to AntNet can significantly reduce the average packet delay and jitter by about 11.8% and 46% respectively in transient regime. It can also improve the throughput in link failure state by about 25% and decrease the packet loss rate by 27% and 39% before and during link failure respectively. Moreover, the proposed approach outperforms the WSP and SWP algorithms, where network efficiency in saturated load gains an 8.39% improvement while the satisfaction of delay and jitter constraints for time critical applications are also achieved. In saturated load, efficiency is a very important parameter and a few changes in it can result in high performance of network to delivery of data. 2014-01 Thesis NonPeerReviewed text en http://psasir.upm.edu.my/id/eprint/76128/1/ITMA%202014%2014%20IR.pdf Baygi, Maassoumeh Javadi (2014) Multi agent quality of service routing based on scheme ant colony optimization algorithm. Doctoral thesis, Universiti Putra Malaysia. Routing (Computer network management) - Mathematical models Computer network protocols |
| spellingShingle | Routing (Computer network management) - Mathematical models Computer network protocols Baygi, Maassoumeh Javadi Multi agent quality of service routing based on scheme ant colony optimization algorithm |
| title | Multi agent quality of service routing based on scheme ant colony optimization algorithm |
| title_full | Multi agent quality of service routing based on scheme ant colony optimization algorithm |
| title_fullStr | Multi agent quality of service routing based on scheme ant colony optimization algorithm |
| title_full_unstemmed | Multi agent quality of service routing based on scheme ant colony optimization algorithm |
| title_short | Multi agent quality of service routing based on scheme ant colony optimization algorithm |
| title_sort | multi agent quality of service routing based on scheme ant colony optimization algorithm |
| topic | Routing (Computer network management) - Mathematical models Computer network protocols |
| url | http://psasir.upm.edu.my/id/eprint/76128/ http://psasir.upm.edu.my/id/eprint/76128/1/ITMA%202014%2014%20IR.pdf |