Is there a simple way of extending the Stochastic Growth Model with one extra control variable?
The brent_max optimizer only takes in one free variable?
Would be interesting to add variable which represent the decision of having capital placed in risky-asset vs capital in fixed income.
Guess, the first change will be to set the Cobb-Douglas function to a linear production function.
And just optimize the value function with respect to consumption and asset-share?
But note that we use these kinds of routines because they work in jit-compiled functions (those with @jit at the start), which helps with execution speed.
If you don’t mind a little bit more execution time, you can drop the jit-compilation and then pick any optimization routine from SciPy — there are lots.
I just updated the optimal growth lecture on the Python side to use @jitclass, which was fun. The code is a bit cleaner now. But it still needs to be checked for typos, so apologies in advance if you find some.