Add two new terms, change organization.

[luciansmith]

• Add 'ODE Solver', and move 'CVODE Solver' and 'Livermore solver' to be children of it.
• Add 'parameters for' for specifying more-specific parameters for more-specific algorithms.

[luciansmith]

This is my attempt to implement the suggestions from HARMONY and described at

https://groups.google.com/g/sed-ml-discuss/c/nIJOYsPGEag

However, there may well be KiSAO conventions I broke, so if someone could take a look, that'd be great.

[luciansmith]

Both Euler and Runge Kutta are themselves children of the 'one-step method' group: should I move the entire group to be children of ODE Solver? Here's the whole list, with the hierarchy missing because it didn't cut-and-paste properly, but at least everything is there:

'one-step method' 'Euler method' 'Euler backward method' 'Euler forward method' 'Euler-Maruyama method' 'Richardson extrapolation based method' 'Bader-Deuflhard method' 'Bulirsch-Stoer algorithm' 'Runge-Kutta based method' 'Butcher-Kuntzmann method' 'embedded Runge-Kutta method' 'Dormand-Prince 8(5,3) method' 'embedded Runge-Kutta 5(4) method' 'Cash-Karp method' 'Numerical Recipes in C "quality-controlled Runge-Kutta" method' 'Dormand-Prince method' 'Fehlberg method' 'hierarchical Fehlberg method' 'Higham-Hall method' 'embedded Runge-Kutta Prince-Dormand (8,9) method' 'Merson method' 'Verner method' 'Zonneveld method' 'explicit fourth-order Runge-Kutta method' 'explicit Runge-Kutta method of order 3(2)' 'Gauss-Legendre Runge-Kutta method' 'Gill method' 'Hammer-Hollingsworth method' 'Heun method' IMEX 'implicit 4th order Runge-Kutta method at Gaussian points' 'implicit midpoint rule' 'Lobatto method' 'midpoint method' 'modified midpoint method' 'partitioned Runge-Kutta method' 'Radau method' 'Rosenbrock method' 'Numerical Recipes in C "stiff" Rosenbrock method' 'semi-implicit midpoint rule' 'stochastic Runge-Kutta method' 'stochastic second order Runge-Kutta method'

[matthiaskoenig]

Makes sense. These are all ODE solvers.

On Mon, May 22, 2023, 16:32 Lucian Smith @.***> wrote:

Both Euler and Runge Kutta are themselves children of the 'one-step method' group: should I move the entire group to be children of ODE Solver? Here's the whole list, with the hierarchy missing because it didn't cut-and-paste properly, but at least everything is there:

'one-step method' 'Euler method' 'Euler backward method' 'Euler forward method' 'Euler-Maruyama method' 'Richardson extrapolation based method' 'Bader-Deuflhard method' 'Bulirsch-Stoer algorithm' 'Runge-Kutta based method' 'Butcher-Kuntzmann method' 'embedded Runge-Kutta method' 'Dormand-Prince 8(5,3) method' 'embedded Runge-Kutta 5(4) method' 'Cash-Karp method' 'Numerical Recipes in C "quality-controlled Runge-Kutta" method' 'Dormand-Prince method' 'Fehlberg method' 'hierarchical Fehlberg method' 'Higham-Hall method' 'embedded Runge-Kutta Prince-Dormand (8,9) method' 'Merson method' 'Verner method' 'Zonneveld method' 'explicit fourth-order Runge-Kutta method' 'explicit Runge-Kutta method of order 3(2)' 'Gauss-Legendre Runge-Kutta method' 'Gill method' 'Hammer-Hollingsworth method' 'Heun method' IMEX 'implicit 4th order Runge-Kutta method at Gaussian points' 'implicit midpoint rule' 'Lobatto method' 'midpoint method' 'modified midpoint method' 'partitioned Runge-Kutta method' 'Radau method' 'Rosenbrock method' 'Numerical Recipes in C "stiff" Rosenbrock method' 'semi-implicit midpoint rule' 'stochastic Runge-Kutta method' 'stochastic second order Runge-Kutta method'

— Reply to this email directly, view it on GitHub https://github.com/SED-ML/KiSAO/pull/110#issuecomment-1557330133, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAG33OSWZXNBEADPWZLX26DXHN2JHANCNFSM6AAAAAAYIKICV4 . You are receiving this because you commented.Message ID: @.***>

[luciansmith]

OK, changes made! Thanks again for the comments.

[luciansmith]

OK, going ahead and merging this. Do let me know if anyone needs something more.