I have used both - started with Repast and currently
with Mason - and will offer some opinion and observation:
-- Repast has a larger code base
If you are interested in using the software mostly
as is, you may find Repast more attractive from a
social science perspective. However, it generally
is not difficult to transfer modules from Repast
to Mason (IMHO).
-- Mason has a cleaner design for extension
While this is true from a number of perspectives,
it is most notable wrt Repast when considering
visualization. The Repast model is untenable (IMHO)
if you plan any sort of sophisticated visualization.
Mason generally allows you free reign with regards
to Java graphics access and control, while for Repast
the model is deeply embedded (ingrained?) with lots
of unmanageable dependencies.
-- Other Observations
I have found Mason easier to work with and more natural
to the way I think about approaching and managing
interacting agents in constrained dynamic environments.
I have also found Mason easier to extend, and not just
with visualization - to the point I have built a
specialized simulator supporting dynamic interconnection
environments with interacting agents on top of Mason
for studying decentralized reasoning in ad hoc environments.
I don't know enough about Jason to offer an opinion on ease of
integration or use, though in my experience the trouble often lies
with the software to be brought in rather than with Mason (whereas
with Repast I often had trouble bringing it together with other
software). However, a larger issue may lie with SACI if that is
GŁnther Greindl wrote:
> I have surveyed the field of ABM simulation software,
> and have finally decided I will either settle for Repast or
> Mason - both appeal to me: Repast because it is specifically
> designed to meet Social Simulation needs and seems to has
> a big following in the socsim community, and Mason because of it's
> design (on the code level). (lean and mean ;-)
> Has anybody tested both platforms and can offer some insight
> on the differences (I'm interested in _opinions_ :-))?
> I would need it to model cultural diffusion processes, my approach
> stems from viewing human societies as complex adaptive systems
> and special interest will be invested in modeling the cultural processes
> (either with cognitive agents, or "memetic" data structures or something
> in that direction).
> A good starting point seems to me the sugarscape model (I also want
> to include resources and competition between the agents).
> The agents should be adaptive in the long run (read: learning ability) -
> although this goes somewhat contrary to the keep it simple paradigm.
> There is an implementation for Repast, it is also listed on the Mason site
> as a project but I did not find a download.
> Basically, what I'm interested in is a very stable platform with an open
> community, and an ease of integrating other technologies (for instance
> BDI with Jason)
> Best Regards,