An Approach towards Stochastic Ambients

(I also caught a bit of the previous talk on semantic subtyping for π. It looks great! Also very codenameesque… anyway, to the talk at hand:)

WWW is the application considered - probabilities of event occurrences are important there (probability of, what, packet loss? server error?), as in bioambients.

The exponential distribution function is 1-e-rt

  • if you use the exp. dist. fn, you have a CTMC
  • and it’s memoryless

Ambients get decorated with a parameter η which is a scaling factor, controlling the rate of reaction inside the ambient. Also actions in, out etc. get decorated with a rate just like stochastic-π.

[How does that work with bioambients? Would the η be like volume there? What happens if you try to stuff a huge ambient inside a tiny one (in bio)?]

There’s no implementation yet. This is still very preliminary work.