June 2005


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

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

Print Reply
MASON Multiagent Simulation Toolkit <[log in to unmask]>
Yiwen Zhang <[log in to unmask]>
Mon, 27 Jun 2005 03:18:59 -0400
MASON Multiagent Simulation Toolkit <[log in to unmask]>
text/plain (55 lines)
Hi, dear multi-agent software experts,

I am new to MASON. I have a very interesting multi-agent modeling and
simulation of online community. Unfortunately I recently lost my java
programming, and need to rebuild the program ASAP. I was wondering whether
MASON is suitable for my model. The description of my model is as following.
(The description is kind of long. I would appreciate it a lot if you can
finishing reading it and give me some suggestions).

Basically, in my model, I have agents and public place.
1) each participant of the online community is modeled as an agent, and it
has a lot of parameters describing their learning capabilities,  social
capabilities, communities's attactiveness to the agent, etc. These
parameters are randomly assigned when the agents are generated, and will
dynamically change over the time.
2) Agents are not always in the community. Each agent periodically visits
the community once it is generated (based on some time distribution), and
when the attractivenss of the community to the agent decrease to zero, it
will never come back again (it's like it died)
3) every time an agent visits the community, agent may initiate messages,
read others' messages, and reply others' messages.The detail of the
activities, such as how many message to generate, which message to reply,
are controlled by the learning and social parameters of the agent.

Public place:
1) The online community is simplified as a public place, where the agents
can post messages, read messages, and reply others' messages. Each message
is represented by a set of parameters (such as author, age, topic, etc)
determined by the parameters  of the agent who posted the message. Public
place is represented by a list of messages. Messages are not necessary
stored forever, and they will be cleared up after a period of time.
2) The interactions among agents are achieved through the public place. When
agents post messages, it potentially will affect other agents if other
agents read the message and even respond to them. There is no direct
agent-to-agent interaction.


I have a set of interesting questions. For example, how do different
compositions of population (different types of agents with different
learning capabilities and social capability) lead to the different growth
pattern of the online community?

Do you think MASON is good for my model? If it is, could you please provide
me some suggestions on what components or steps I should take when
implementing my model in MASON. If it is not, do you have any suggestions on
other software?

Thanks a ton!