Loading...
Graph Signal Processing-Based Resource Allocation in Fiber-Optic Networks
Zarei Ardestani, Mohammad Hossein | 2022
57
Viewed
- Type of Document: M.Sc. Thesis
- Language: Farsi
- Document No: 56493 (05)
- University: Sharif University of Technology
- Department: Electrical Engineering
- Advisor(s): Hadi, Mohammad; Pakravan, Mohammad Reza
- Abstract:
- With the advancement of technology, the volume of network traffic has also increased dramatically. The optical network, as the main part of the backbone communication system, plays a significant role in the transmission of network traffic. Due to the limited network resources, an efficient resource allocation in optical networks is required to increase network performance and reduce costs. To solve resource allocation optimization problems, an effective tool is needed to provide a suitable solution in a reasonable time. Resource allocation problems are usually developed on a graph, which is derived from network topology or other properties of the network. As a result, graph-based tools seem to be useful and lucrative for resource allocation problems. Graph signal processing is one of the tools that has been used a lot in recent years to solve various problems. Considering the capabilities of graph signal processing and its high ability to model various problems, it is expected that graph signal processing is a suitable tool for solving optical resource allocation problems. Therefore, in this research, we intend to use graph signal processing to solve resource allocation problems in optical networks, specifically routing and wavelength assignment, and routing and spectrum assignment problems. In this regard, after introducing the problems and concepts related to optical networks, we describe the graph signal processing tool. Then, by modeling network resources in the form of graph signals, we present methods based on graph signal processing to solve the mentioned resource allocation problems. In these methods, the fitness between traffic requests and network resources is evaluated with the help of graph signal processing concepts. In order to demonstrate the optimal resource management of these methods, we perform numerous simulations. The simulation results show that the proposed methods reduce request blocking probability by 10-25% compared to the common benchmark methods
- Keywords:
- Optical Networks ; Resources Allocation ; Graph Signal Processing ; Optical Fibre Network ; Optimum Allocation ; Network Traffic