Using symbolic mathematics to find equations of motion

[Bruno_Mendes]

I am working with a reduced-order mathematical model for the motion of a offshore platform.
The deduction of equations of motion requires a lot of symbolic computation as differentiation of time-dependent variables and inversion of symbolic matrices. I was using Wolfram Mathematica with this purpouse. What packages can be used for this type of algebraic manipulation in Julia?

[ChrisRackauckas]

@Bruno_Mendes

The deduction of equations of motion requires a lot of symbolic computation as differentiation of time-dependent variables and inversion of symbolic matrices. I was using Wolfram Mathematica with this purpouse. What packages can be used for this type of algebraic manipulation in Julia?

You might want to check out this workshop which does this in full using the universal differential equation methodology:

JuliaCon 2020 | Doing Scientific Machine Learning (SciML) With Julia

DataDrivenDiffEq.jl is the library to look at for doing this:

https://datadriven.sciml.ai/dev/sparse_identification/sindy/

and the example mixes neural networks into the training methodology can be found here:

https://github.com/ChrisRackauckas/universal_differential_equations/blob/master/DelayLotkaVolterra/VolterraExp.jl