Post date: Oct 2, 2014 9:12:32 AM
Due to late cancellation of our scheduled speaker, Stuart Rossiter (University of Southampton) gave a talk on how software engineering best-practice can relate to simulation design (across modelling paradigms). Abstract below:
Simulation Design: Trans-Paradigm Best-Practice from Software Engineering
or 'What have software engineers ever done for us simulators?'
Stuart Rossiter (University of Southampton, Care Life Cycle project)
[This is material that is hopefully due in a forthcoming issue of the Journal of
Artificial Societies & Social Simulation (JASSS).]
Although simulations are ultimately pieces of software, software design
best-practice often tends to be ignored in the simulation literature. I discuss
some potential reasons why, and try to show that there *are* important patterns
of design which apply to simulations across paradigms.
This involves introducing an 'idealised' software architecture (what software
engineering would call a 'reference architecture') for simulations which
incorporates some best-practices. I show how this relates to some existing
simulation toolkits (MASON and AnyLogic), and then how applying it works in
practice on an existing multi-paradigm model of health and social care.