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
julia>parse(BigFloat,"0.546136459064141770613483785351029091974067802242010192661484687977676021081218637004123196030567166414008915304513669838826570884623515686") inu(0) # true value from mathematicatrue
However, this is mostly luck: the convergence criteria doesn't actually control the decay in the tail. To do this properly (and thereby have apriori guaranteed success) we would need to modify the convergence check at
while (k ≤min(m+M,A_dim) || BLAS.nrm2(M,yp,1) > tolerance )
to properly bound the tail.
The text was updated successfully, but these errors were encountered:
dlfivefifty
changed the title
ApproxFun+IntervalSets: make ODE solving really rigorous
ApproxFun+IntervalArithmetic: make ODE solving really rigorous
May 7, 2019
The following code returns successfully for solving Airy's equation (the extra definitions are work arounds for bugs):
It happens to have worked:
However, this is mostly luck: the convergence criteria doesn't actually control the decay in the tail. To do this properly (and thereby have apriori guaranteed success) we would need to modify the convergence check at
ApproxFunBase.jl/src/Caching/matrix.jl
Line 122 in 7dd9ef0
to properly bound the tail.
The text was updated successfully, but these errors were encountered: