Mismatch on Figure 7.38 about worlddynamics.jl HOT 7 CLOSED

We should test what we get if we solve the system like them, i.e. with a simple Euler method with fixed step size of 5 years.

from worlddynamics.jl.

The solution returned by the Euler solver is noisy.

The changes have been made under solvesystems.jl in the following manner (corresponding to the first plot below) :

function solve(system::ODESystem, timespan; solver=Euler())
    sys = structural_simplify(system)

    prob = ODEProblem(sys, [], timespan)
    sol = ModelingToolkit.solve(prob, solver, dt = 5)

    return sol
end

Fixed step size of 5 :

Fixed step size of 50 :

from worlddynamics.jl.

Yes, @piluc and I also verified this independently.
Now @piluc and @paulobruno are looking into this.

from worlddynamics.jl.

There must be an error in the book, since our Figure agrees with the one of pyworld3.

from worlddynamics.jl.

To illustrate what @natema mentioned, here is the comparison of fig 7.38.

World3	WorldDynamics.jl	PyWorld3

from worlddynamics.jl.

Since the general agreement is that our code is now returning the correct plot, I'm closing this issue.

from worlddynamics.jl.

For future reference, the same mismatch is present in Figure 7.39 as well.

from worlddynamics.jl.