Loading ...
Sorry, an error occurred while loading the content.
 

RE: [decentralization] Talk/Event Reminder: Tue Jul 23, 2002, 2:0 0pm -- Distributed Rational Decision Making (fwd)

Expand Messages
  • cefn.hoile@bt.com
    Lucas, I find myself stuck into this problem (cooperation vs competition) very regularly, since the work we are doing in DIET, (our Peer to Peer Multi-Agent
    Message 1 of 1 , Jul 29, 2002
      Lucas,

      I find myself stuck into this problem (cooperation vs competition) very
      regularly, since the work we are doing in DIET, (our Peer to Peer
      Multi-Agent Toolkit), often uses Evolutionary Computation as a means of
      agent adaptivity. Consequently, the global behaviour we are after had better
      be a consequence of self-interest, as agents evolve new ways to exploit
      their environment for their own benefit without human control.

      Like many DHTs, SWAN depends crucially on the conformant behaviour of its
      participants, so I agree with your definition as 'cooperative problem
      solving'. However, depending upon your interpretation of 'unstructured
      networks', SWAN is similar to both a DHT and an unstructured network. It is
      like a DHT in that individuals adopt responsibility for some part of the
      addressing load. It is an unstructured network in that resources do not need
      to be relocated in order to maintain scalability.

      Cefn

      >>>>>>>

      Lucas Wrote....

      I'm forwarding this relative to the thread on A* and Swan. This is the
      relevant bit from the text:

      *******************************

      In cooperative distributed problem solving, the system designer imposes an
      interaction protocol and a strategy (a mapping from state history to
      action; a way to use the protocol) for each agent. The main question is
      what social outcomes follow given the protocol and assuming that the
      agents use the imposed strategies.

      On the other hand, in multiagent systems, the agents are provided with an
      interaction protocol, but each agent will choose its own strategy. A
      self-interested agent will choose the best strategy for itself, which
      cannot be explicitly imposed from outside. Therefore, the protocols need
      to be designed using a noncooperative, strategic perspective: the main
      question is what social outcomes follow given a protocol which guarantees
      that each agent's desired local strategy is best for that agent---and thus
      the agent will use it.

      *******************************


      It strikes me that distributed hash tables are what this document calls
      "cooperative distributed problem solving", and that unstructured networks
      are what is called "multiagent systems".

      - Lucas
    Your message has been successfully submitted and would be delivered to recipients shortly.