New to julia, apologies if this is trivial. I am struggling to work out how to pass parameters to an ODE.
In short, I have two systems: one 7N dimensional (with N in the hundreds or thousands) and a 7-D one. I am not even at the point of running the 7N-D version, but in planning to do this: parameters of that one are both vectors with N entries and N\times N-matrices.
Every run of the ODE, I choose different values of the parameters and need to run the 7-D version first, picking the relevant parameters. So I cannot manually prepare a vector of parameters (and I most definitely cannot do in the 7N-D version).
What I have this far is a Dict() with a load of different fields (“keys”/“values” pairs?). I do something like
p_tmp = Dict()
for curr_param in ["eta", "xi", "d", "epsilon", "gamma", "beta"]
p_tmp[curr_param] = p[curr_param][p["idx_positive_IC"]]
end
to pick up the relevant parameters for the 7-D version. Then set up the solver and call.
prob = ODEProblem(SLLLDUU_rhs, IC, (tspan[1], tspan[end]), p_tmp)
sol = solve(prob)
I get the error
ERROR: ArgumentError: This problem does not support symbolic maps with `remake`, i.e. it does not have a symbolic origin. Please use `remake`with the `p` keyword argument as a vector of values (paying attention toparameter order) or pass `interpret_symbolicmap = false` as a keyword argument
It is my understanding that I should be using remake, but I don’t understand the purpose of that function and I don’t like using things I don’t understand. Or I could transform to a NamedTuple using something like
p_tmp = NamedTuple([pair for pair in p_tmp])
But that fails (ERROR: TypeError: in typeassert, expected Symbol, got a value of type String). Or I could use ComponentArrays, but I am not sure how that would scale up to entries that are matrices. (Actually, not sure about that for NamedTuples either.)
Anyway, any recommendations on what to do? Thanks!!