python-dolfin-adjoint 2019.1.0 Automatic differentiation library
python-dolfin-adjoint is a solver of differential equations associated with a governing system and a functional of interest. Working from the forward model the solver automatically derives the discrete adjoint and tangent linear models. These additional models are key ingredients in many algorithms such as data assimilation, optimal control, sensitivity analysis, design optimisation and error estimation. The dolfin-adjoint project provides the necessary tools and data structures for cases where the forward model is implemented in
fenics or firedrake.