DAEO Solver with Optima Tracking

Solve equations of the form $$ x(0, x) = x_0\ x'(t, x, y) = f(t, x, y)\ y(t) ∈ \argmin_y h(x, y) $$ by tracking all possible local optima $y_i$ of the objective function $h$.

Build Guide

• This project requires some of C++23 (std::ranges).

• This project requires fmt/libfmt, Eigen 3.4+ and Boost (boost interval depends on boost::limits...)

• NAG's dco/c++ CMake scripts cannot locate dco/c++ installations with the new product code. In this case, you'll need to make sure dco.hpp and libdcoc.a are in your build path by default.

• This project uses our fork of Boost's interval, which updates some parts of the library to C++11 and fixes some bugs caused by use with Eigen 3.4. It is provided as a submodule here.

• Output logs are, by default, placed in data/out, which is created when CMake configures.

To-Do List for AD2024

• Investigate the use of the implicit function theorem to find $\partial_x y^k$.
• Handle the emergence and disappearance of local optima.
• Implement a (justifiable) heuristic for re-running global optimization.