You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
We could use RangeEnclosures for computing the abstraction.
It has an interface with a list of solvers so if our abstraction based solvers are parametrized by a AbstractEnclosureAlgorithm then they automatically support any alogrithm implementing the RangeEnclosures API !
It's still only for scalar output but in their paper, they say vector output is planned, let's keep an eye on it !
If what we still want to use what we are currently doing by default, we could implement it as a solver and have it as default solver.
I also just added a PR for better support of vector outputs, at least for some of the solvers, in JuliaReach/RangeEnclosures.jl#175. But it may require some discussions before that gets in.
We could use RangeEnclosures for computing the abstraction.
It has an interface with a list of solvers so if our abstraction based solvers are parametrized by a
AbstractEnclosureAlgorithm
then they automatically support any alogrithm implementing theRangeEnclosures
API !It's still only for scalar output but in their paper, they say vector output is planned, let's keep an eye on it !
If what we still want to use what we are currently doing by default, we could implement it as a solver and have it as default solver.
cc @schillic @mforets @lucaferranti
The text was updated successfully, but these errors were encountered: