Scribd Logo
Upload

Welcome to Scribd!

  • 31 views

    Why Using Game Theory in Communication Networks

    Uploaded by

    nangaayissi
    Game theory concepts can be used to analyze communication networks where there are competitive actors whose individual actions impact the network performance and other actors. The document discusses how networking has evolved from centralized to decentralized systems, and introduces basic game theory concepts like strategic games, utility functions, Nash equilibria, and applications to problems like pricing, congestion control, routing, and competition among network providers.

    Copyright:

    © All Rights Reserved
    Download as DOC, PDF, TXT or read online from Scribd

    Why Using Game Theory in Communication Networks

    Uploaded by

    nangaayissi
    31 views6 pages
    Game theory concepts can be used to analyze communication networks where there are competitive actors whose individual actions impact the network performance and other actors. The document discusses how networking has evolved from centralized to decentralized systems, and introduces basic game theory concepts like strategic games, utility functions, Nash equilibria, and applications to problems like pricing, congestion control, routing, and competition among network providers.

    Original Description:

    ffgjhjhkk::lk:kk

    Copyright

    © © All Rights Reserved

    Available Formats

    DOC, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Game theory concepts can be used to analyze communication networks where there are competitive actors whose individual actions impact the network performance and other actors. The document discusses how networking has evolved from centralized to decentralized systems, and introduces basic game theory concepts like strategic games, utility functions, Nash equilibria, and applications to problems like pricing, congestion control, routing, and competition among network providers.

    Copyright:

    © All Rights Reserved
    Download as DOC, PDF, TXT or read online from Scribd
    Download as doc, pdf, or txt
    31 views6 pages

    Why Using Game Theory in Communication Networks

    Uploaded by

    nangaayissi
    Game theory concepts can be used to analyze communication networks where there are competitive actors whose individual actions impact the network performance and other actors. The document discusses how networking has evolved from centralized to decentralized systems, and introduces basic game theory concepts like strategic games, utility functions, Nash equilibria, and applications to problems like pricing, congestion control, routing, and competition among network providers.

    Copyright:

    © All Rights Reserved
    Download as DOC, PDF, TXT or read online from Scribd
    Download as doc, pdf, or txt
    You are on page 1of 6

    Why using Game Theory in communication networks?

    Introduction of game theory concepts

    Patrick   Maill´e   and   Bruno   Tuffin  Institut   Mines­Telecom/Telecom


    Bretagne,   Inria   Rennes   Bretagne­Atlantique  IRISA   Seminar   on
    Network Economics, May 31, 2012

    Outline
    Introduction: the (economic) evolution of networks 

    Basic concepts of game theory

    Pricing and congestion/demand control Interdomain issues

    5 Summary

    From centralization to decentralization

    Networking has switched from the centralized telephone network to
    the decentralized Internet (scalability reason)

    . Decentralization (or deregulation) is a key factor

    .In such a situation:

     From the decentralization, there is a general envisaged/advised
    behavior
     But   each  selfish  user   can   try   to   modify   his   behavior   at   his
    benefits and at the expense of the network performance.

    Example: TCP configuration

     How   to  analyze  this,  and how to  control  and  prevent  such  a


    thing? It is the purpose of non­cooperative game theory.

    Competitive actors: not only users

    The   Internet   has   also   evolved   from   an   academic   to   a
    commercial   network   with   providers   in   competition   for
    customers and services.

    As a consequence, users are not the only  competitive  actors, but


    also 

     network providers: several providers propose the same type of
    network access;
     Applications/service providers/content providers: the same type
    of  application   can   be proposed  by several entities (ex: search
    engines...)
     Platforms/technologies:   you   may   access   the   Internet   from
    ADSL,   WiFi,   3G,   WiMAX,   LTE...All   those   interacting   actors
    have to be considered

    Why changing the pricing scheme?

     Increase of Internet traffic due to
    o increasing number of subscribers
    o more and more demanding applications.

     Congestion is a consequence, with erratic QoS.
     Increasing   capacity   difficult   if   not   impossible   in   access
    networks (last mile problem)
     Also,  flat  rate  pricing  is unfair and does  not  allow  service
    differentiation.

    Subject of debate...

    But new contexts require new economic paradigms.

    Convergence: requires new pricing offers

     Convergence   of   technologies   and   services:   all   services   (web


    browsing, telephony, television) can be used on all technologies
    (Fixed, WiFi, 3G, LTE...).
    o How   to   charge   fairly   and   efficiently   those   different
    technologies, with their different characteristics?
    o New technology: new issues to solve.
     Would   it   be   possible   to   propose   a   pricing   scheme   involving
    several technologies at the same time?
     Marketing point of view from operators: propose grouped offers
    (bundles)   to   attract   customers   to   services   they   would   not
    consider otherwise.

    A new issue: dealing with competition among providers

     Most works on pricing are dealing with a monopoly, but
    o Competition   forces   providers   to   decide   about   prices   and
    offers depending on competitors’ ones.
    o Some a priori promising pricing schemes may not resist to
    competition.
     Sometimes   providers   even   operate   on   different   technologies
    (Fixed, WiFi, 3G, LTE...), or on several simultaneously.
     Also,   impact   of   competition   on   capacity   investment?   Do   they
    have interest in investing (especially when congestion pricing is
    used)?

    Sending   end­to­end   traffic   through   other   (selfish)


    nodes/networks

     Here not a direct competition for customers, but providers have
    to   pay   other   domains   for   forwarding   their   traffic   and   ensure
    end­to­end delivery (similar problems arise in ad hoc networks).
     How   to   design   a   self­managed   network,   with   proper   pricing
    incentives to forward traffic?
     Still unsolved: what are the optimal strategies of operators, in
    order to propose the best investments, in terms of :
    1. Investment on capacity: bandwidth for a domain or mobile
    network...
    2. Investment on products: new services.
    3. Investment on technology: new link between two domains,
    new base station, new WiFi hotspot...
    Regulation: is it required?

     Free market can lead to an \inefficient" mechanism.
     Regulation   can   enforce   providers   to   drive   to   the   proper
    situation.
     Ex: to enforce providers to reduce retention time and authorize
    churn.
     New regulation/political issue: network neutrality
    o Network providers want to win on both sides: to charge users
    but also content providers, or degrade their services.
    o They   do   not   want   application   providers   not   associated   to
    them to congest their network.
    o Political   debate:   all   players   should   be   allowed   the   same
    access.
     Actors are then not free to do whatever we want.

    Basic concepts of game theory

    What it changes
     While   before   optimization   was   the   tool   for   routing,   QoS
    provisionning, interactions between players has to be taken into
    account.
     Game  theory:   distributed  optimization:  individual  users make
    their  own   decisions.   "Easier"  than  to  solve  NP­hard problems
    (approximation).
     We   need   to   look   at   a   stable   point   (Nash   equilibrium)   for
    interactions.
     Tools   used   before   in   Economics,   Transportation...
    and has recently appeared in telecommunications.
     We may have paradoxes (Braess paradox) that can be studied
    that way.
     A   way   to   control   things:   to   introduce   pricing
    incentives/discouragements (TBC).

    Typical networking applications


     P2P networks: a node tries to benefit from others, but limits its
    available resource (free riding)
     Grid   computing:   same   issue,   try   to   benefit   from   others’
    computing power, while limiting its own contribution.
     Routing   games:   each   sending   node   tries   to   find   the   route
    minimizing delay, but intermediate links are shared with other
    flows (interactions).
     Ad   hoc   networks:   what   is   the   incentive   of   nodes   to   forward
    traffic   of   neighbors?   If   no   one   does,   no   traffic   is   successfully
    sent.
     Congestion   control   game   (TCP...):  why   reducing   your   sending
    rate when congestion is detected? 
     Power control in wireless networks: maximizing your power will
    induce a better QoS, but at the expense of others’ interferences.
     Transmission   games   (WiFi...):  if   collision,   when   to   resubmit
    packets?

    Basic definitions

     Game   theory:   set   of   tools   to   understand   the   behavior   of


    interacting decision makers or players.
     Classical   assumption:   players   are  rational:   they   have   well­
    defined objectives,  and they take into account the behavior of
    others.
     In   this   course:  strategic   or   normal   games,   players   play
    (simultaneously) once and for all.
     There are also branches called:
    o extensive games, for which players play sequentially;
    o repeated  games  for which they can  change their choices
    over time;
    o Bayesian games, evolutionary games...

    General modeling tools

     Interactions of players through network performance. Tools:
    o queueing analysis or
    o signal processing.
     The   action   of   a   player   has   an   impact   on   the   output   of   other
    players, and therefore on their own strategies.
     They all have to play strategically.
     Each   player   i   (user   or   provider)   represented   by   its   utility
    function ui(x) representing quantitatively its level of satisfaction
    (in monetary units for instance) when actions profile is x = (x i)i,
    where xi denotes the action of player i.

    Strategic Games
     A strategic game Γ consists of:
    o A finite set of players, N.
    o A set Ai of actions available to each player.
    o For each player a utility function, (payoffs) characterizing
    the gain/utility from a state of the game.
     Players   make   decisions   independently,   without   information
    about the choice of other players.
     We note Γ 
     For   two   players:   description   via   a   table,   with   payoffs
    corresponding to the strategic choices of users:

    You might also like

    pFad - Phonifier reborn

    Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

    Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


    Alternative Proxies:

    Alternative Proxy

    pFad Proxy

    pFad v3 Proxy

    pFad v4 Proxy