Skip to content

ABM Software Comparison

James Hawley edited this page Aug 12, 2015 · 1 revision

Table of desired functions + status in NetLogo/MASON

Table showing everything we want the ABM to do (e.g. each agent has a spread that is an f(t), 3D models) and whether it's "achieved" (green), "in progress" (yellow), "not possible" (red) for both MASON and NetLogo.

(To edit table )

Criterion NetLogo MASON
Different kinds of agent for different cells Achieved Achieved
Control over # connections Acheived Not as easy as we like
Dynamic connections between cells. Easy
Killing agents (apoptosis) Achieved Achieved
Arbitrary number of states (for MOI) In progress (well supported) Achieved
ODEs (complex rules) within each agent Probably not Probably
Time behaviour – arbitrary function of time since infection Not easily - may have to insert Java function Almost done
Easy to use? Yes It's Java
Speed Good; number of cells limited Good, my Macbook Air can get ~40000 cells
Nice visualizations? Achieved -- decent. Hard to get really good. Ok, but charts are a pain...can use outside software
3D models? Maybe Lots of work

Install/Setup Instructions

The installation/setup for NetLogo is fairly simple: press the Download button on this page (I don't think you actually need to fill out any of the fields): https://ccl.northwestern.edu/netlogo/download.shtml

To edit the code for the model, 'open' the file from inside the NetLogo program

Parameters for both software

(I don't think "software" is a countable noun)

Table parameters for each software (though we might do some to combine them) with value, description, source and a note on how the software interprets the parameter. If you have realistic values for anything, should have level of information as the CRISPRi table from last year (linked at bottom of e-mail).

Note about parameters in NetLogo -- many parameters are (in current implementation) set with sliders in the GUI. This makes it really easy to experiment with them, but there are some issues -- namely, all parameter values are implicitly interdependent with the size of the time step. There's only one size of "tick" supported by NetLogo (afaik); the speed slider just changes how quickly the the visual display evolves. If you want to simulate a finer timestep (say 1/2 as big), you may have to manually adjust all parameters -- cut the linear ones by 1/2, the multiplicative ones by sqrt 2 (or ln 2 or whatever) and so on.

Clone this wiki locally