MASON-INTEREST-L Archives

June 2015

MASON-INTEREST-L@LISTSERV.GMU.EDU

Options: Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Subject:
From:
Russell Thomas <[log in to unmask]>
Reply To:
MASON Multiagent Simulation Toolkit <[log in to unmask]>
Date:
Mon, 29 Jun 2015 11:59:05 -0700
Content-Type:
text/plain
Parts/Attachments:
text/plain (105 lines)
Thanks for your replies: Christian, Ernesto, Joey, and Siggy.

Yes, the Strategy pattern does fit my needs better and addresses some
problems I was running into.  Thanks for the links and references.  I found
a good tutorial in Java:
http://www.javaworld.com/article/2074195/swing-gui-programming/strategy-for-
success.html

Russ


On 6/29/15, 4:22 AM, "Christian Meyer" <[log in to unmask]> wrote:

> Using an entity component system (ECS) might be suitable for you. It is
> similar to the strategy pattern but any combination of "strategies" is
> possible and also saving state along with them.
> 
> Each agent is an entity which is nothing more than a set of components
> (representing state). A factory class is used to create entities with
> the needed component sets. During each step systems (representing logic)
> will update the entity by changing its components. For example, a
> ReproductionSystem will look for a certain subset of components needed
> for reproduction and only update agents who have this subset. To give an
> agent the ability to reproduce, just add the needed components and you
> are done.
> 
> This is a simple ECS which I changed for MASON. Some work was needed to
> integrate it into the schedule.
> https://github.com/adamgit/Entity-System-RDBMS-Beta--Java-
> 
> Best regards,
> Christian
> 
> On 29.06.2015 03:23, Russell Thomas wrote:
>> Is anyone using the ³Factory² design pattern to construct agents at run
>> time? I am about halfway down this path and Iıd love to learn from
>> anyone else who has experience with this approach. (See ³Difficulties²
>> below)
>> 
>> Requirements
>> ------------------
>> 
>> The setting is modeling teams and organizations engaged in design (of
>> artificial products), innovation (in how they do design), and
>> qualitative learning (i.e. learning new skills, not just getting better
>> at existing skills).  In other words, Iım modeling endogenous innovation
>> involving genuine novelty.  My agents are (somewhat) cognitive and
>> adaptive, particularly in their ability to acquire or develop novel
>> capabilities during the course of a simulation run. Also Iıd like to
>> experiment with a wide variety of initial conditions regarding agent
>> types and capabilities.  In essence, what Iım looking for is agents that
>> are dynamic bundles of capabilities (i.e. resources, action rules, etc.)
>> rather than fixed capabilities as is the norm in most ABMs.
>> 
>> Approach
>> -------------
>> These requirements suggest the use of a ³Factory² design pattern (see
>> https://en.wikipedia.org/wiki/Factory_method_pattern for general
>> description).
>> 
>> During start( ), the /AgentFactory /is an object that takes as input
>> specifications for the agents (i.e. list of capabilities) and also other
>> key objects (i.e. environment objects such as /SimState/).  The
>> /AgentFactory /class includes various pre-built templates and objects to
>> produce an agent object of the specified type.  (I have a large library
>> of capability and processing objects.)  Actions to be performed by agent
>> at each time step is specified by a special-purpose language that
>> describe sequences of tasks and their interdependencies. (Execution is
>> by a simple O-O stack machine.)  Each task execution records its
>> results, and these results can be concatenated and logged in a file.
>>   Essentially, agents can ³explain² what they did and why each step.
>>   This is useful for both debugging and also as experimental results.
>> 
>> Clearly, this approach does not result in fast execution and would not
>> be suitable for a large population of agents and long simulation runs.
>> My simulations will involve tens or at most a hundred agents, so it
>> wonıt be a problem for me.
>> 
>> Difficulties
>> --------------
>> Where Iım running into difficulty is in *software design decisions*: 1)
>> how general-purpose to make the code, and 2) how to avoid unnecessary
>> complications.  The advantage of making it fully general-purpose is that
>> it will (later) reduce the design and debug time for specific ABM
>> designs.  The disadvantage is that it takes more time now for thinking,
>> planning, coding, and testing (e.g. debug capabilities for the
>> special-purpose stack machine).
>> 
>> In pursuit of general functionality, I am probably making the software
>> design too complicated.  It would be nice to see what other people have
>> done and maybe learn from their experience.
>> 
>> Also, I want to use built-in capabilities of Java and also available
>> Java libraries.  For example, when I get to the point where agent
>> specifications will come from a text file, Iıll probably use ANTLR to
>> parse my special-purpose task language.  But I might be unknowingly
>> ³recreating the wheel² in many cases by creating my own classes for
>> various low-level functions.
>> 
>> Thanks,
>> 
>> Russell Thomas
>> Computational Social Science
>> George Mason University

ATOM RSS1 RSS2