DiffEqDevTools.jl

DiffEqDevTools.jl copied to clipboard

``` julia> solve(prob, Tsit5()).dense true julia> solve(prob, Tsit5(), dense=false).dense false julia> TestSolution(solve(prob, Tsit5())).dense true julia> TestSolution(solve(prob, Tsit5(), dense=false)).dense true ``` This is unexpected to me, and might be a bug.

nathanaelbosch

correct constructTanakaKasugaYamashitaYazaki6A to D and constructMikkawyEisa

1

Instead of https://github.com/SciML/DiffEqDevTools.jl/blob/260b90d8634042dd3df2c194db43643fbe5162d0/src/ode_tableaus.jl#L2469 constructTanakaKasugaYamashitaYazaki6D should return return(ExplicitRKTableau(A,c,α,6,adaptiveorder=5,αEEst=αEEst,fsal=true)) --- Instead of https://github.com/SciML/DiffEqDevTools.jl/blob/260b90d8634042dd3df2c194db43643fbe5162d0/src/ode_tableaus.jl#L2554 constructTanakaKasugaYamashitaYazaki6C should return return(ExplicitRKTableau(A,c,α,6,adaptiveorder=5,αEEst=αEEst,fsal=true)) --- Instead of https://github.com/SciML/DiffEqDevTools.jl/blob/260b90d8634042dd3df2c194db43643fbe5162d0/src/ode_tableaus.jl#L2639 constructTanakaKasugaYamashitaYazaki6B should return return(ExplicitRKTableau(A,c,α,6,adaptiveorder=5,αEEst=αEEst,fsal=true)) --- Instead of https://github.com/SciML/DiffEqDevTools.jl/blob/260b90d8634042dd3df2c194db43643fbe5162d0/src/ode_tableaus.jl#L2724 constructTanakaKasugaYamashitaYazaki6A should return...

emca1

Error reduction function

1

@devmotion 's Quorum Sensing problem is an interesting case where the "full error" is not what we really want, but a more refined function for how to calculate the error....

ChrisRackauckas

Fixed L2 error for complex valued problems

7

Currently, L2 error calculations in DiffEqDevTools.jl use `sqrt(recursive_mean(vecvecapply((x) -> float(x) .^ 2, sol - timeseries_analytic)))` the issue is if the solution is complex this returns the sum of the square...

henhen724