Re: Organization Self-Design

[Rob Kewley]

Thomas A Ottaway wrote:
> 
> Hello,
> 
> I would like to model several forms of organizational self-desig.  I would
> like to have an agent, under predetermined conditions, decompose into two
> or more agents, complete their respective subtasks, and then recombine.
> Esentially, I want to repeat some work done by Ishida, Yokoo, and Gasser
> for different purposes.  Can Swarm handle this type of
> decomposition/composition of agents.  Thanks for any assistance.
> 

Thomas:

        I am sure swarm can handle the task you wish.  There is no built-in
mechanism for your specific task.  However, the simplest implementation
would be to use a dynamic schedule for your simulation and allow the
agents to schedule events for themselves.  For example, if one of an
agent's recurring tasks is "update," then at the end of each update, it
would schedule the next update for itself at the appropriate
"updateInterval".  As for decomposing, you may never want to destroy the
original agent.  When the agent decomposes, it should create the two (or
more) subordinate agents then schedule the necessary actions for those
agents on the dynamic schedule.  The original agent simply stops
scheduling events for itself until it gets messages from the subordinate
agents which state they have completed their sub tasks.  When the
original agent receives all such messages, it simply begins scheduling
events for itself again.  Hope this helps.  I am sure there are other
ways to do what you want, but this is one way.

Rob Kewley

                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.
                  ==================================