i'm thinking about computing the zonotopic oa by default, and add a plot recipe for taylor model flowpipes that uses it.
note that if an invariant is passed, we already compute it inside _is_intersection_empty(Ri, X) if intersection_method is a ZonotopeEnclosure() (the default)
overapproximate(<taylor model flowpipe>, ::Zonotope) should check if that approx has been computed.
this change would be aligned to making the library more user-friendly: one gets a first plot for a nonlinear ODE without actually knowing about overapproximations, zonotopes, etc.
this can be done by returning a tuple,
function _is_intersection_empty(X::AbstractReachSet, Y::LazySet, ::ZonotopeEnclosure)
Z = overapproximate(X, Zonotope)
return (Z, _is_intersection_empty(set(Z), Y))
end
i'm thinking about computing the zonotopic oa by default, and add a plot recipe for taylor model flowpipes that uses it.
note that if an invariant is passed, we already compute it inside
_is_intersection_empty(Ri, X)
ifintersection_method
is aZonotopeEnclosure()
(the default)overapproximate(<taylor model flowpipe>, ::Zonotope)
should check if that approx has been computed.this change would be aligned to making the library more user-friendly: one gets a first plot for a nonlinear ODE without actually knowing about overapproximations, zonotopes, etc.
this can be done by returning a tuple,