Implement simple baseline value function solutions with DiffEqOperators and compare

@PooyaFa @ajozefiak

We are getting closer to having a full library for dealing with differential operators and composition, but there are likely bugs we need to work through. So @PooyaFa I was hoping you could work with Adam to use the DiffEqOperators.jl to implement some of the test cases we have built up. A good one to start with is https://quantecon.github.io/SimpleDifferentialOperators.jl/stable/examples/#Solving-HJBE-with-constant-drifts-1

My suggestion is the following:

[x] Can you guys work together in a PR for now so we don't bring down the whole package?
[ ] @ajozefiak , decide how you want to manage dependencies on this. For now, you could put DiffEqOperators into the test dependencies for https://github.com/QuantEcon/SimpleDifferentialOperators.jl/blob/master/Project.toml and keeping the Manifest in check
[x] Maybe create a subfolder /test/DiffEqOperators/ where you put the tests with the comparisons
[x] Start with the exact example in https://quantecon.github.io/SimpleDifferentialOperators.jl/stable/examples/#Solving-HJBE-with-constant-drifts-1 except drop N. In general, I think that you should use the smallest N possible so it is easier to track what is going on in the matrices. I think starting with 3 interior nodes and 5 on the extension is the right way to go.
[x] Then do the same thing with positive drift instead of negative drift (i.e. in the SDO code, don't forget to switch the direction of L₁₋bc to the correct L₁+bc
[x] Always compare two things in the tests: First, the concreteizaiton and second the A \ x solution, which may use a different solver method when we hook iterative methods into the setup.
[x] After that, do https://quantecon.github.io/SimpleDifferentialOperators.jl/stable/examples/#Solving-HJBE-with-state-dependent-drifts-1 with state varying drifts to make sure the automatic upwinding is working.

QuantEcon / SimpleDifferentialOperators.jl