111 Fast Spectrum Allocation in Coordinated Dynamic Spectrum Access Based Cellular Networks Anand Prabhu Subramanian, Himanshu Gupta, Samir R. Das State.

Slides:



Advertisements
Similar presentations
February 20, Spatio-Temporal Bandwidth Reuse: A Centralized Scheduling Mechanism for Wireless Mesh Networks Mahbub Alam Prof. Choong Seon Hong.
Advertisements

Impact of Interference on Multi-hop Wireless Network Performance Kamal Jain, Jitu Padhye, Venkat Padmanabhan and Lili Qiu Microsoft Research Redmond.
Minimum Clique Partition Problem with Constrained Weight for Interval Graphs Jianping Li Department of Mathematics Yunnan University Jointed by M.X. Chen.
Fast Algorithms For Hierarchical Range Histogram Constructions
A Prior-Free Revenue Maximizing Auction for Secondary Spectrum Access Ajay Gopinathan and Zongpeng Li IEEE INFOCOM 2011, Shanghai, China.
1 Reinforced Tabu Search (RTS) for Graph Colouring Daniel Porumbel PhD student (joint work with Jin Kao Hao and Pascale Kuntz) Laboratoire d’Informatique.
1 Routing and Wavelength Assignment in Wavelength Routing Networks.
Wavelength Assignment in Optical Network Design Team 6: Lisa Zhang (Mentor) Brendan Farrell, Yi Huang, Mark Iwen, Ting Wang, Jintong Zheng Progress Report.
1 NP-Complete Problems. 2 We discuss some hard problems:  how hard? (computational complexity)  what makes them hard?  any solutions? Definitions 
End-to-End Fair Bandwidth Allocation in Multi-hop Wireless Ad Hoc Networks Baochun Li Department of Electrical and Computer Engineering University of Toronto.
Online Social Networks and Media. Graph partitioning The general problem – Input: a graph G=(V,E) edge (u,v) denotes similarity between u and v weighted.
The number of edge-disjoint transitive triples in a tournament.
Complexity 16-1 Complexity Andrei Bulatov Non-Approximability.
Introduction to Approximation Algorithms Lecture 12: Mar 1.
Wireless Mesh Networks 1. Architecture 2 Wireless Mesh Network A wireless mesh network (WMN) is a multi-hop wireless network that consists of mesh clients.
1 Minimum-energy broadcasting in multi-hop wireless networks using a single broadcast tree Department of Computer Science and Information Engineering National.
P-Center & The Power of Graphs A part of the facility location problem set By Kiril Yershov and Alla Segal For Geometric Optimizations course Fall 2010.
Beneficial Caching in Mobile Ad Hoc Networks Bin Tang, Samir Das, Himanshu Gupta Computer Science Department Stony Brook University.
Recent Development on Elimination Ordering Group 1.
Energy-Efficient Target Coverage in Wireless Sensor Networks Mihaela Cardei, My T. Thai, YingshuLi, WeiliWu Annual Joint Conference of the IEEE Computer.
CS541 Advanced Networking 1 Spectrum Sharing in Cognitive Radio Networks Neil Tang 3/23/2009.
Cache Placement in Sensor Networks Under Update Cost Constraint Bin Tang, Samir Das and Himanshu Gupta Department of Computer Science Stony Brook University.
NP-Complete Problems Reading Material: Chapter 10 Sections 1, 2, 3, and 4 only.
On the Construction of Energy- Efficient Broadcast Tree with Hitch-hiking in Wireless Networks Source: 2004 International Performance Computing and Communications.
Online Data Gathering for Maximizing Network Lifetime in Sensor Networks IEEE transactions on Mobile Computing Weifa Liang, YuZhen Liu.
The community-search problem and how to plan a successful cocktail party Mauro SozioAris Gionis Max Planck Institute, Germany Yahoo! Research, Barcelona.
Backtracking Reading Material: Chapter 13, Sections 1, 2, 4, and 5.
1 Algorithms for Bandwidth Efficient Multicast Routing in Multi-channel Multi-radio Wireless Mesh Networks Hoang Lan Nguyen and Uyen Trang Nguyen Presenter:
Graph Coloring.
Domain decomposition in parallel computing Ashok Srinivasan Florida State University COT 5410 – Spring 2004.
CHAMELEON : A Hierarchical Clustering Algorithm Using Dynamic Modeling
Fast Spectrum Allocation in Coordinated Dynamic Spectrum Access Based Cellular Networks Anand Prabhu Subramanian*, Himanshu Gupta*,
MAXIMIZING SPECTRUM UTILIZATION OF COGNITIVE RADIO NETWORKS USING CHANNEL ALLOCATION AND POWER CONTROL Anh Tuan Hoang and Ying-Chang Liang Vehicular Technology.
Efficient Gathering of Correlated Data in Sensor Networks
Network Aware Resource Allocation in Distributed Clouds.
1 11 Subcarrier Allocation and Bit Loading Algorithms for OFDMA-Based Wireless Networks Gautam Kulkarni, Sachin Adlakha, Mani Srivastava UCLA IEEE Transactions.
Design Techniques for Approximation Algorithms and Approximation Classes.
June 21, 2007 Minimum Interference Channel Assignment in Multi-Radio Wireless Mesh Networks Anand Prabhu Subramanian, Himanshu Gupta.
Quasi-static Channel Assignment Algorithms for Wireless Communications Networks Frank Yeong-Sung Lin Department of Information Management National Taiwan.
Batch Scheduling of Conflicting Jobs Hadas Shachnai The Technion Based on joint papers with L. Epstein, M. M. Halldórsson and A. Levin.
Maximum Network Lifetime in Wireless Sensor Networks with Adjustable Sensing Ranges Cardei, M.; Jie Wu; Mingming Lu; Pervaiz, M.O.; Wireless And Mobile.
DISCERN: Cooperative Whitespace Scanning in Practical Environments Tarun Bansal, Bo Chen and Prasun Sinha Ohio State Univeristy.
1 Steiner Tree Algorithms and Networks 2014/2015 Hans L. Bodlaender Johan M. M. van Rooij.
Advanced Spectrum Management in Multicell OFDMA Networks enabling Cognitive Radio Usage F. Bernardo, J. Pérez-Romero, O. Sallent, R. Agustí Radio Communications.
1 Short Term Scheduling. 2  Planning horizon is short  Multiple unique jobs (tasks) with varying processing times and due dates  Multiple unique jobs.
Course: Logic Programming and Constraints
On Reducing Broadcast Redundancy in Wireless Ad Hoc Network Author: Wei Lou, Student Member, IEEE, and Jie Wu, Senior Member, IEEE From IEEE transactions.
CS774. Markov Random Field : Theory and Application Lecture 02
1 11 Channel Assignment for Maximum Throughput in Multi-Channel Access Point Networks Xiang Luo, Raj Iyengar and Koushik Kar Rensselaer Polytechnic Institute.
CSE 589 Part VI. Reading Skiena, Sections 5.5 and 6.8 CLR, chapter 37.
Efficient Computing k-Coverage Paths in Multihop Wireless Sensor Networks XuFei Mao, ShaoJie Tang, and Xiang-Yang Li Dept. of Computer Science, Illinois.
Joint Power and Channel Minimization in Topology Control: A Cognitive Network Approach J ORGE M ORI A LEXANDER Y AKOBOVICH M ICHAEL S AHAI L EV F AYNSHTEYN.
Resource Allocation in Hospital Networks Based on Green Cognitive Radios 王冉茵
ICS 353: Design and Analysis of Algorithms Backtracking King Fahd University of Petroleum & Minerals Information & Computer Science Department.
On the Ability of Graph Coloring Heuristics to Find Substructures in Social Networks David Chalupa By, Tejaswini Nallagatla.
ICS 353: Design and Analysis of Algorithms NP-Complete Problems King Fahd University of Petroleum & Minerals Information & Computer Science Department.
Impact of Interference on Multi-hop Wireless Network Performance
Cohesive Subgraph Computation over Large Graphs
Greedy & Heuristic algorithms in Influence Maximization
Design and Analysis of Algorithm
Maximal Independent Set
Computability and Complexity
ICS 353: Design and Analysis of Algorithms
Buddhikot, M.M. Ryan, K. Lucent Technol. Bell Labs.;
Department of Information Management National Taiwan University
Algorithms for Budget-Constrained Survivable Topology Design
ICS 353: Design and Analysis of Algorithms
Survey on Coverage Problems in Wireless Sensor Networks
Locality In Distributed Graph Algorithms
Presentation transcript:

111 Fast Spectrum Allocation in Coordinated Dynamic Spectrum Access Based Cellular Networks Anand Prabhu Subramanian, Himanshu Gupta, Samir R. Das State University of New York at Stony Brook Milind M. Buddhikot Alcatel-Lucent Bell Labs DySPAN 2008

222 Outline Introduction Model Description Maximum Demands Serviced Dynamic Spectrum Access (Max-Demand DSA) Minimum Interference Dynamic Spectrum Access (Min-Interference DSA) Performance Evaluation Conclusions

3 Introduction (1) Measurement studies have shown that the cellular spectrum is highly utilized but the spectrum utilization varies dramatically over space and time [3–5] Cellular networks will continue to evolve to higher access speeds and therefore, will require larger amount of spectrum However, releasing more spectrum using current long- term command-and-control model of spectrum licensing is a flawed approach [3] Shared Spectrum, Inc. (2006). [Online]. Available: [4] M. A. McHenry, P. A. Tenhula, D. McCloskey, D. Roberson, and C. Wood, “Chicago Spectrum Occupancy Measurements and Analysis and a Long-term Proposal,” in First Workshop on Technology and Policy for Accessing Spectrum (TAPAS 2006), August [5] T. Kamakaris, M. M. Buddhikot, and R. Iyer, “A Case for Coordinated Dynamic Spectrum Access in Cellular Networks,” in Proceedings of IEEE DySPAN05, Baltimore,Maryland, November 2005.

4 Introduction (2) Buddhikot et al. proposed a concept of Coordinated Dynamic Spectrum Access (CDSA) for cellular networks to enable capacity-on-demand services [5, 7, 8] A centralized spectrum broker coordinates access to spectrum in a given region and assigns short term spectrum leases to competing radio infrastructure providers One of the main challenges in building such brokers is the design of fast spectrum allocation algorithms [7] M. M. Buddhikot, P. Kolodzy, S. Miller, K. Ryan, and J. Evans, “DIMSUMnet: New directions in wireless networking using coordinated dynamic spectrum access,” in Proceedings of IEEE WoWMoM 2005, June [8] M. M. Buddhikot and K. Ryan, “Spectrum management in coordinated dynamic spectrum access based cellular networks,” in Proceedings of IEEE DySPAN05, Baltimore,Maryland, November 2005

5 Contributions Formulate the spectrum allocation problem as two optimization problems Max-Demand DSA: with the objective of maximizing the overall spectrum demands satisfied among various base stations such that no two interfering base stations that belong to different radio infrastructure providers are assigned the same channels Min-Interference DSA: with the objective of minimizing the overall interference in the network when all the demands of the base stations are satisfied Propose a graph construct called interference graph that captures conflict relationships between transmitters of various radio infrastructure providers that co-exist in a region Develop constant factor approximation algorithms for the Max- Demand DSA problem and Min-Interference DSA problem

6 Model Description (1) We designate the smallest amount of contiguous spectrum that can be requested via CDSA as a channel of C units If the broker manages a spectrum band of B units, it can dynamically allocate K = B/C channels Each spectrum demand request for a base station is specified as a range between d min and d max channels We assume a batched spectrum request processing model the spectrum demands received in a time window of τ units are grouped and processed together the allocated spectrum is used in subsequent time windows

7 Model Description (2) In this model, a part of the spectrum, designated as the Coordinated Access Band (CAB), is meant to be dynamically shared under the control of a spectrum broker Each region R, which is under the control of a spectrum broker can have a number of base stations (nodes) owned by several Radio Infrastructure Providers (RIPs) The Wireless Service Providers (WSPs) who offer wireless services such as voice, data etc. to the end users are customers of these RIPs and may use different RIPs in different regions and at different times The network elements such as the Radio Network Controllers (RNCs) that control the base stations aggregate the end user demands and generate a spectrum demand request to the spectrum broker

8 Interference Graph The networks of various RIPs in the region R controlled by the spectrum broker are modeled as a weighted undirected graph called the interference graph G =(V,E) each base station is represented by a node in the graph There is an edge (i, j) ∈ E between nodes i and j, if the base stations represented by them belong to different RIPs and interfere with each other Each edge (i, j) ∈ E has a weight p ij associated with it which is the penalty when nodes i and j are assigned the same channels

9 Notations

10 Maximum Demands Serviced Dynamic Spectrum Access (Max-Demand DSA) The objective of Max-Demand DSA is to maximize the overall demand serviced such that no two base stations belonging to different service providers that interfere with each other are assigned same channels

11 Relationship with Maximum K-Colorable Induced Subgraph (Max K-CIS) Problem Definition of Max K-CIS: Given a graph G =(V,E) and an integer K, find a K-colorable subgraph of G with the maximum number of vertices Reduction assume the minimum demands of each node is 0 given the interference graph G(V,E), create a new graph G max =(V max, E max ) such that for each node i ∈ V, we create d max (i) copies of it in V max and form a clique among those nodes for each edge (i, j) ∈ E, add to E max an edge from each copy of node i to each copy of node j color a node in G max means we are servicing one demand of a base station

12 Maximum Independent Set Problem (Max-IS) Definition of Max-IS: find a set of vertices of maximum cardinality such that no two vertices have an edge between them It can be shown that approximating the Max K-CIS problem is as hard as approximating the Max-IS problem for any fixed value of K [14] A solution to the Max-K-CIS problem can be obtained by repeating the Max-IS algorithm K times removing the nodes in independent set formed from the graph in every iteration [14] D. S. Hochbaum, Approximation algorithms for NP-hard problems. Boston, MA, USA: PWS Publishing Co., 1997

13 δ-degree Bounded Graph The Max-K-CIS problem is hard to approximate in general graphs we use a δ-degree bounded graph to model the interference graph Definition A graph G =(V,E) is said to be δ-degree bounded, if the maximum node degree of any node in G is less than or equal to δ Considering the sparse nature of deployment of base stations, a δ- degree bounded graph capture the characteristics of a realistic cellular network quite well

14 Max-IS algorithm Algorithm 1. Pick a node i ∈ V such that the maximum independent set in the induced subgraph in the neighborhood of i is minimum among all nodes 2. Add i to the solution IS and remove i and all its neighbors from V 3. Repeat step 1 and 2 until all vertices in V are removed from the graph

15 Max-Demand DSA Algorithm (1) Phase I: Given the interference graph G =(V,E), create a new graph G min =(V min,E min ) each node i ∈ V, we create d min (i) copies of it in V min and form a clique among those nodes For each edge (i, j) ∈ E, we add an edge from each copy of node i to each copy of node j to E min Try to color the nodes of graph G min using K colors by solving the Max-K-CIS problem in graph G min the Max-K-CIS problem can be solved by repeating Max-IS algorithm K times

Max-Demand DSA Algorithm (2) Phase II: Add extra copies (d max (i) − d min (i)) of each node i ∈ V to the already colored graph G min to form the new graph G max Solve the Max-K-CIS problem in G max to color as many extra vertices as possible using the K colors 16

17 Minimum Interference Dynamic Spectrum Access (Min-Interference DSA) The objective of Min-Interference DSA is to minimize the overall interference in the network when all the demands (d max ) of the base stations are serviced

18 Relationship with Max-K-Cut Problem Definition of Max-K-Cut: Given a graph G =(V, E), find a K-partitioning of the vertex set V, such that the number of edges that have their endpoints in different partitions is maximized The weighted Max-K-Cut problem partition the vertex set such that the sum of the weights of the edges whose endpoints are in different partitions is maximized Reduction assume the maximum demand of each node to be 1 our problem boils down to assigning one of the K colors to each node such that the sum of the weights (p ij ’s) of the monochromatic edges is minimized monochromatic edges: edges with endpoints assigned the same color

19 Multi-Color Max-K-Cut Problem Definition of Multi-Color Max-K-Cut: Given the weighted interference graph G = (V,E) with demands d max (i) for each node i ∈ V and the total number of colors K Assign d max (i) different colors to each node i such that the sum of the weights of the non-monochromatic edges is maximized i.e. maximize where F(i) is the set of colors assigned to i and F(j) is the set of colors assigned to j

20 R k Algorithm The R k (random) algorithm: For each node i, randomly pick d max (i) different colors from the available K colors and assign them to node i Each color is expected to be in F(i) with a probability of d max (i)/K The probability of any particular color being in F(i) as well as F(j) is d max (i)d max (j)/K 2 The expected value of (|F(i) ∩ F(j)|) is d max (i)d max (j)/K The expected value of the R k solution is

Tabu Search Algorithm for Min- Interference DSA (1) Tabu search [20] based heuristic starts with the random solution obtained by algorithm R k improves the solution to get a better solution for the Min- Interference DSA problem Algorithm Start with a random initial solution F 0 wherein each node i ∈ V is assigned to d max (i) different random colors in κ In the lth iteration (l ≥ 0), we create the next solution F l+1 in the sequence (from F l ) 21 [20] A. Hertz and D. de Werra, “Using tabu search techniques for graph coloring,” Computing, vol. 39, no. 4, 1987

Tabu Search Algorithm for Min- Interference DSA (2) The lth Iteration First, we generate a certain number (say, r) of random neighboring solutions of F l a random neighboring solution of F l is generated by picking a random vertex i ∈ V and a color in F l (i) and changing it to a random color in (K−{F l (i)}) generate 100 neighboring solutions in each iteration Pick the neighboring solution with the lowest network interference as the next solution F l+1 Termination Keep track of the best solution F best seen so far Terminate the algorithm when the maximum number of allowed iterations have passed without any improvement in I(F best ) 22

Simulation Environment Graph Parameters 1000 nodes in the network randomly assigned to 10 service providers Each node has a transmission range of 150m Two nodes have an edge between them, if they belong to different service providers and are within 300m from each other We generated graphs of different densities by randomly placing the 1000 nodes in a fixed area of size 23

Performance of Max-Demand DSA We used four sets of demands In the first set, the minimum demand of each node was randomly picked from 1 to 10 and the maximum demand for each node was randomly picked from 10 to 20 Similarly we used the values (20,40),(30,60) and (40,80) for the other three sets of demands 24 maximum node degree 10

Performance of Min-Interference DSA We used four sets of demands In the first set of demands, each node picks a value randomly from 1 to 10 for d max Similarly we used the values 20,30,40 for the other three sets of demands 25 maximum node degree 10

Conclusions We reported two formulations of the spectrum allocation problem as two optimization problems: first with the objective of maximizing the overall number of demands (Max-Demand) satisfied among the various base stations the second with the objective of minimizing the overall interference in the network (Min-Interference) when all the demands of the base stations are satisfied We showed that the optimization problems are NP- hard and designed efficient algorithms to solve them 26