[Dcopolis-development] DCOPolis Status Update

[Evan A. S. <ev...@su...>]

Things I've done in the past two days:

  * Completely re-designed the Platform interface
  * Created a `Networkless' platform that runs in simulated time on a
    single machine, but does not use MATES for simulating the network.
    This platform is equivalent to simulating a bunch of nodes on a
    fixed network.
  * Finished creating all of the necessary classes in the `problem'
    package.
  * Created a `RandomGraphColoringProblem' DCOP.
  * Finished all of the classes for the `algorithm' package.
  * Implemented a really stupid algorithm that I call
    `DistributedHillClimbing' (it's not optimal, but it works).
  * Created a bunch of visualization APIs that let the platform and
    problem visualize their current state.
  * Created a base GUI for DCOPolis that lets one select a problem,
    algorithm, and platform.

Right now, the networkless platform, distributed hill climbing
algorithm, and random graph coloring problem all seem to work.  The
TCPPlatform seems to be broken, but I haven't worked on it very much.

Now we're in a state that we can at least start coding up "real" DCOP
algorithms like ADOPT and test them against the RandomGraphColoring
problem using the NetworklessPlatform.

To get the current code running:

$ svn co https://dcopolis.svn.sourceforge.net/svnroot/dcopolis dcopolis
$ cd dcopolis
$ ant compile
$ java -cp lib/sefirs.jar:bin org.dcopolis.DCOPolis

    - Evan