Simultaneous Routing and Resource Allocation in Wireless Networks Mikael Johansson Signals, Sensors and Systems, KTH Joint work with Lin Xiao and Stephen.

Slides:



Advertisements
Similar presentations
3.6 Support Vector Machines
Advertisements

EE384y: Packet Switch Architectures
1 UNIT I (Contd..) High-Speed LANs. 2 Introduction Fast Ethernet and Gigabit Ethernet Fast Ethernet and Gigabit Ethernet Fibre Channel Fibre Channel High-speed.
Rethinking Internet Traffic Management From Multiple Decompositions to a Practical Protocol Martin Suchara in collaboration with: J. He, M. Bresler, J.
Cognitive Radio Communications and Networks: Principles and Practice By A. M. Wyglinski, M. Nekovee, Y. T. Hou (Elsevier, December 2009) 1 Chapter 11 Information.
Cognitive Radio Communications and Networks: Principles and Practice By A. M. Wyglinski, M. Nekovee, Y. T. Hou (Elsevier, December 2009) 1 Chapter 10 User.
Cognitive Radio Communications and Networks: Principles and Practice By A. M. Wyglinski, M. Nekovee, Y. T. Hou (Elsevier, December 2009) 1 Chapter 12 Cross-Layer.
Mission-based Joint Optimal Resource Allocation in Wireless Multicast Sensor Networks Yun Hou Prof Kin K. Leung Archan Misra.
Energy-Efficient Distributed Algorithms for Ad hoc Wireless Networks Gopal Pandurangan Department of Computer Science Purdue University.
1 Multi-Channel Wireless Networks: Capacity and Protocols Nitin H. Vaidya University of Illinois at Urbana-Champaign Joint work with Pradeep Kyasanur Chandrakanth.
and 6.855J Cycle Canceling Algorithm. 2 A minimum cost flow problem , $4 20, $1 20, $2 25, $2 25, $5 20, $6 30, $
Tuesday, March 5 Duality – The art of obtaining bounds – weak and strong duality Handouts: Lecture Notes.
Interference Avoidance and Control Ramki Gummadi (MIT) Joint work with Rabin Patra (UCB) Hari Balakrishnan (MIT) Eric Brewer (UCB)
Scalable Routing In Delay Tolerant Networks
Multipath Routing for Video Delivery over Bandwidth-Limited Networks S.-H. Gary Chan Jiancong Chen Department of Computer Science Hong Kong University.
1 Outline relationship among topics secrets LP with upper bounds by Simplex method basic feasible solution (BFS) by Simplex method for bounded variables.
Solve Multi-step Equations
MIMO Broadcast Scheduling with Limited Feedback Student: ( ) Director: 2008/10/2 1 Communication Signal Processing Lab.
Two-Market Inter-domain Bandwidth Contracting
Utility Optimization for Event-Driven Distributed Infrastructures Cristian Lumezanu University of Maryland, College Park Sumeer BholaMark Astley IBM T.J.
Particle Swarm Optimization (PSO)
1 Column Generation. 2 Outline trim loss problem different formulations column generation the trim loss problem master problem and subproblem in column.
Detection Chia-Hsin Cheng. Wireless Access Tech. Lab. CCU Wireless Access Tech. Lab. 2 Outlines Detection Theory Simple Binary Hypothesis Tests Bayes.
Shadow Prices vs. Vickrey Prices in Multipath Routing Parthasarathy Ramanujam, Zongpeng Li and Lisa Higham University of Calgary Presented by Ajay Gopinathan.
The Weighted Proportional Resource Allocation Milan Vojnović Microsoft Research Joint work with Thành Nguyen Microsoft Research Asia, Beijing, April, 2011.
Hash Tables.
MAT 205 F08 Chapter 12 Complex Numbers.
COMP 482: Design and Analysis of Algorithms
1 Capacity analysis of mesh networks with omni or directional antennas Jun Zhang and Xiaohua Jia City University of Hong Kong.
College of Engineering Capacity Allocation in Multi-cell UMTS Networks for Different Spreading Factors with Perfect and Imperfect Power Control Robert.
1 COPYRIGHT © 2011 ALCATEL-LUCENT. ALL RIGHTS RESERVED. On the Capacity of Wireless CSMA/CA Multihop Networks Rafael Laufer and Leonard Kleinrock Bell.
نیمسال اوّل افشین همّت یار دانشکده مهندسی کامپیوتر مخابرات سیّار (626-40) ظرفیت انتقال اطلاعات.
Capacity of wireless ad-hoc networks By Kumar Manvendra October 31,2002.
Chapter 6 The Mathematics of Diversification
Routing and Congestion Problems in General Networks Presented by Jun Zou CAS 744.
1 Scheduling Crossbar Switches Who do we chose to traverse the switch in the next time slot? N N 11.
Dantzig-Wolfe Decomposition
PSSA Preparation.
Introduction to Ad-hoc & Sensor Networks Security In The Name of God ISC Student Branch in KNTU 4 th Workshop Ad-hoc & Sensor Networks.
© 2007 Levente Buttyán and Jean-Pierre Hubaux Security and Cooperation in Wireless Networks Chapter 11: Wireless operators in.
Mani Srivastava UCLA - EE Department Room: 6731-H Boelter Hall Tel: WWW: Copyright 2003.
Chapter 5 The Mathematics of Diversification
Delay Analysis and Optimality of Scheduling Policies for Multihop Wireless Networks Gagan Raj Gupta Post-Doctoral Research Associate with the Parallel.
Peter Key, Laurent Massoulie, Don Towsley Infocom 07 presented by Park HoSung 1 Path selection and multipath congestion control.
13-Optimization Assoc.Prof.Dr. Ahmet Zafer Şenalp Mechanical Engineering Department Gebze Technical.
New Opportunities for Load Balancing in Network-Wide Intrusion Detection Systems Victor Heorhiadi, Michael K. Reiter, Vyas Sekar UNC Chapel Hill UNC Chapel.
1 EL736 Communications Networks II: Design and Algorithms Class8: Networks with Shortest-Path Routing Yong Liu 10/31/2007.
Decomposable Optimisation Methods LCA Reading Group, 12/04/2011 Dan-Cristian Tomozei.
ACCESS Group meeting Mikael Johansson Novel algorithms for peer-to-peer optimization in networked systems Björn Johansson and Mikael.
1 Cross-Layer Design for Wireless Communication Networks Ness B. Shroff Center for Wireless Systems and Applications (CWSA) School of Electrical and Computer.
EE 685 presentation Optimization Flow Control, I: Basic Algorithm and Convergence By Steven Low and David Lapsley Asynchronous Distributed Algorithm Proof.
1 Algorithms for Bandwidth Efficient Multicast Routing in Multi-channel Multi-radio Wireless Mesh Networks Hoang Lan Nguyen and Uyen Trang Nguyen Presenter:
DaVinci: Dynamically Adaptive Virtual Networks for a Customized Internet Jennifer Rexford Princeton University With Jiayue He, Rui Zhang-Shen, Ying Li,
Optimization Flow Control—I: Basic Algorithm and Convergence Present : Li-der.
Scalable Multi-Class Traffic Management in Data Center Backbone Networks Amitabha Ghosh (UtopiaCompression) Sangtae Ha (Princeton) Edward Crabbe (Google)
Minimax Open Shortest Path First (OSPF) Routing Algorithms in Networks Supporting the SMDS Service Frank Yeong-Sung Lin ( 林永松 ) Information Management.
EE 685 presentation Utility-Optimal Random-Access Control By Jang-Won Lee, Mung Chiang and A. Robert Calderbank.
DaVinci: Dynamically Adaptive Virtual Networks for a Customized Internet Jiayue He, Rui Zhang-Shen, Ying Li, Cheng-Yen Lee, Jennifer Rexford, and Mung.
Simultaneous routing and resource allocation via dual decomposition AUTHOR: Lin Xiao, Student Member, IEEE, Mikael Johansson, Member, IEEE, and Stephen.
EE 685 presentation Optimization Flow Control, I: Basic Algorithm and Convergence By Steven Low and David Lapsley.
1 An Arc-Path Model for OSPF Weight Setting Problem Dr.Jeffery Kennington Anusha Madhavan.
1 Slides by Yong Liu 1, Deep Medhi 2, and Michał Pióro 3 1 Polytechnic University, New York, USA 2 University of Missouri-Kansas City, USA 3 Warsaw University.
Resource Allocation in Hospital Networks Based on Green Cognitive Radios 王冉茵
Optimization-based Cross-Layer Design in Networked Control Systems Jia Bai, Emeka P. Eyisi Yuan Xue and Xenofon D. Koutsoukos.
1 Chapter 5 Branch-and-bound Framework and Its Applications.
Joint Routing and Scheduling Optimization in Wireless Mesh Networks with Directional Antennas A. Capone, I. Filippini, F. Martignon IEEE international.
Resource Allocation in Non-fading and Fading Multiple Access Channel
Cross layer design is wireless multi-hop network
Power Efficient Communication ----Joint Routing, Scheduling and Power Control Design Presenter: Rui Cao.
ACHIEVEMENT DESCRIPTION
Presentation transcript:

Simultaneous Routing and Resource Allocation in Wireless Networks Mikael Johansson Signals, Sensors and Systems, KTH Joint work with Lin Xiao and Stephen Boyd, Stanford University

2 About this talk Pedagogical motivation To convey ideas and techniques from distributed convex optimization Technological motivation Wireless ad-hoc networks promising emerging technology Intellectual motivation Will ad-hoc networks deliver the required performance (capacity)? Compute the optimal parameters for a given network configuration Devise simple, distributed protocols that ensure efficient network operation Control-theoretic motivation Distributed resource allocation problems roots of distributed control theory New technological challenges/problems may inspire theoretical advances

3 Outline Motivation System model Optimal routing and resource allocation Example Efficient solution methods Distributed algorithms Conclusions and extensions

4 Motivation: cross-layer optimization Standard (OSI) network model Physical/radio link layer, network layer (routing), transport... Wireless data network Optimal routing of data depends on link capacities Link capacities are determined by resource allocation Efficient operation requires coordination of layers!

5 Disclaimer This talk only considers orthogonal channel models, simple and elegant theory Interference-limited systems require other techniques High signal-to-noise ratio: convex approximation [JXB:03] Low signal-to-noise ratio: scheduling, integer programming [JX:03] In practice, time-varying channels and delays fundamental limitations Very active area of research, many open problems!

6 System model We assume fixed topology fixed coding, modulation and optimize rates, routing & resource allocation We model multiple data flows influence of resource allocation on link capacities local & global resource limits

7 Network topology Directed graph with nodes, links set of outgoing links at node, incoming links at Incidence matrix

8 Network flow model Model average data rates, multiple source/destination pairs Identify flows by destination –source flows flow from node to node –link flows flow on link to node Flow conservation laws

9 Multicommodity network flow Some traditional formulations: fixed, minimize total delay: fixed, maximize total utility:

10 Communications model Capacities determined by resource (power, bandwidth) allocation Communications model Where – is a vector of resources allocated to link, e.g., – is concave and increasing –resource limits local (power at node) or global (total bandwidth) Many (most?) channel models satisfy these assumptions!

11 Example: Gaussian broadcast with FDMA Communication variables: Shannon capacity: Total power, bandwidth constraint on outgoing links

12 Concavity of the capacity formula Claim: capacity formula is jointly concave in powers and bandwidths Proof: its Hessian is negative semi-definite

13 Simultaneous optimization of routing and resource allocation Solution to optimization problem We assume that are convex SRRA is a convex optimization problem, hence readily solved

14 Examples SRRA formulation is very general, includes Maximum utility routing (QoS) Minimum power routing as well as minimum bandwidth, minimax link utilization, etc.

15 Numerical example –50 nodes, 340 links (transmitters) –5 nodes exchange data (i.e., 20 source-destination pairs) –transmitters use FDMA, power limited in each node –goal: maximize network utility

16 Numerical example: details Topology: node locations drawn from uniform distribution on a square two nodes can communicate if distance smaller than threshold source and destination nodes chosen randomly Radio layer properties bandwidth allocation fixed, power constraint at each node quadratic path loss model noise power drawn from uniform distribution Optimization problem has 2060 variables (1720 flows, 340 powers)

17 Numerical example: results Routing to node 1 Aggregate data flows Power allocation Result: Result: 35% improvement over routing w. uniform power allocation Note: log-utility gives diminishing returns, throughput improvements much larger

18 Outline System model Optimal routing and resource allocation Example Efficient solution methods Distributed algorithms Conclusions and extensions

19 Solution methods Small problems readily solved using off-the-shelf software Real-world problems: hundreds of nodes, thousands of links Size proportional to, often large! Can we do better? Can we do it distributedly, in real-time?

20 Duality Primal problem Dual problem Convex duality: optimal values of both problems equal¹ can solve original problem via its dual² Lagrange decomposition: multipliers for critical constraints only Decompose dual into subproblems that are easy to solve Can give very efficient overall optimization

21 Example: Water-filling Consider the following convex optimization problem (equivalent to maximizing weighted total utility) Total power constraint destroys separable structure! Solution approach –introduce Lagrange multiplier for this constraint only –solve dual problem –recover optimal solution

22 Example: Water-filling Dual function Dual problem Solved by adjusting until power constraint becomes tight.

23 Dual decomposition of SRRA Introduce multipliers for capacity constraints only Problem decomposes into Uncapacitated network flow problems (one per commodity) Resource allocation problem (often solved by water-filling)

24 Solving the master dual Master dual problem solved using sub-gradient method step-length parameter, sub-gradient Multipliers decreased when capacity exceeds traffic

25 Convergence of dual method Convergence of dual method vs. number of iterations An alternative approach, the analytic-center cutting plane method, has better convergence requires considerably more computations per iteration appears hard to implement distributedly

26 Economics interpretation Interpret dual variables as price per unit traffic on each link Network layer: Network layer: minimizes network loss + cost of capacities used Radio control layer: Radio control layer: allocates resources to maximize revenue Price updates: Price updates: follow laws of supply and demand

27 Outline System model Optimal routing and resource allocation Example Efficient solution methods Distributed algorithms Conclusions and extensions

28 Distributed algorithms Simplified model for fixed routing The matrix indicates what flows traverse what links Note: relation to TCP Vegas (e.g., [LPW:02]) over wireless links

29 Distributed algorithms Consider a dual approach The first subproblem admits closed-form solution -solved locally by sources, if they know their total path cost The second subproblem solved using water-filling in each node

30 Distributed algorithms Dual problem can be solved using subgradient method Note that multipliers can be computed locally for each link A distributed algorithm: Transport layer: Transport layer: sources maxmimize utility minus resource cost Radio layer Radio layer: nodes allocate resources to maximize revenue Link prices: Link prices: follow supply and demand Convergence follows along the lines of [LL:99].

31 Conclusions Conclusions: Optimal cross-layer coordination in wireless data networks Simultaneous optimization of routing & resource allocation Convex optimization problem, hence readily solved Very efficient solution methods by exploiting structure Distributed methods for adaptive resource allocation More info at