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
I am looking at some SciML problems where, depending on the stability of the ODE, the integrand used w/ HCubature.jl may return Inf. I am noticing inconsistent behavior across hquadrature and hcubature for single and multi-dim domains.
I see similar behavior in Cubature.jl as well, but Cubature.hcubature returns Inf up to dim = 4 and then NaN, while Cubature.pcubature returns Inf up dim = 17 (haven't test passed that).
Is this the expected behavior? I assume this is due to some Inf*0 that is propagating down.
Thanks.
The text was updated successfully, but these errors were encountered:
I = V * (g.w[1]*f₁ + g.w[2]*f₂ + g.w[3]*f₃ + g.w[4]*f₄ + g.w[5]*f₅)
I am happy to put together a PR to address this corner case, but it is unclear to me what the appropriate behavior should be. My initial thought is to short-circuit the algorithm and return Inf if f ever returns Inf.
I am looking at some SciML problems where, depending on the stability of the ODE, the integrand used w/ HCubature.jl may return Inf. I am noticing inconsistent behavior across
hquadrature
andhcubature
for single and multi-dim domains.I see similar behavior in Cubature.jl as well, but
Cubature.hcubature
returns Inf up to dim = 4 and then NaN, whileCubature.pcubature
returns Inf up dim = 17 (haven't test passed that).Is this the expected behavior? I assume this is due to some Inf*0 that is propagating down.
Thanks.
The text was updated successfully, but these errors were encountered: