Skip to content

vboussange/ParametricModels.jl

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

43 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

ParametricModels

Build Status

DifferentialEquations.jl is amazing, but sometimes you want to write an ODE model once for all, and simply play around with the model without bothering further about the details of the numerical solve, etc... If this is the case, ParametricModels.jl is for you!

Getting started

using ParametricModels, UnPack

# Set the model name
@model LogisticGrowth

# Define the ODE system
function (model::LogisticGrowth)(du, u, p, t)
    @unpack r, b = p
    du .=  r .* u .* ( 1. .- b .* u) 
end

tsteps = 1.:0.5:100.5
tspan = (tsteps[1], tsteps[end])
p1 = (r = [0.05, 0.06], b = [0.23, 0.5],)
u0 = ones(2)

# Instantiating the model
mp = ModelParams(p1, 
                tspan,
                u0, 
                BS3(),
                saveat = tsteps, 
                )
model = LogisticGrowth(mp)

# Playing with the model, without worrying about the details!
sol = simulate(model)

# Let's change the parameter values
p2 = (r = [0.05, 0.06], b = [0.4, 0.6],)
sol2 = simulate(model, p = p2)

# Or the initial conditions
u0  = [3., 4.]
sol3 = simulate(model, u0 = u0)

# even cooler: let's change only a subset of the parameters
p2 = (r = [0.05, 0.07],)
sol2 = simulate(model, p = p2)

Defining generic models

The macro @model defines under the hood a struct which inherits AbstractModel. You can write by hand the struct to write generic models, which can e.g. be declined for ODE systems with state variables with arbitrary dimensions, or with exchangeable components.

To define a model, you need the following

Base.@kwdef struct MyModel{EF} <: AbstractModel
    mp::ModelParams
    extra_field::EF #optional
end

And then define the ODE system as usual, where you can use extra_field as you wish.

Make sure to use the macro Base.@kwdef. This defines a constructor MyModel(;mp, extra_field), which is used internally in ParametricModels.jl.

About

An AD friendly wrapper for consistently handling dynamical models

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages