---
license: mit
---

Dataset of sensitivity systems from DiffOpt.jl.

Part of: https://github.com/LearningToOptimize/BatchNLPSolver.jl.

The solution of the linear systems in the dataset gives the sensitivity of the optimal solution of the underlying optimization problem wrt the parameters.

Each file is a different optimization problem. JLD2 files are segregated by (parameter1, ..., parameterm, M, N).
M and N are the LHS and RHS of the linear system respectivelly.

Included problems:
 - Inverse kinematics of a 2dof planar arm.
   - Number of Parameter Variables: 2 (each is a scalar).
   - Parameters Interpretation: End-Point Target.
   - Ref:https://github.com/jump-dev/DiffOpt.jl/blob/master/docs/src/examples/Planar_Arm_Example.jl
 - AC-OPF IEEE 300 Bus case.
   - Number of Parameter Variables: 1 (vector).
   - Parameters Interpretation: Multiplicative Factor Loads.
   - `_corrected`: matrix diagonal values are changed from the original to make the matrix non-singular.
   - Ref:
     - PowerModels: https://github.com/lanl-ansi/PowerModels.jl
     - PGLib: https://github.com/power-grid-lib/pglib-opf