1

I want to solve a quadratic program with mpmath precision. The problem can be stated as a least square with linear equality and inequality constraints, or as quadratic program with linear equality and inequality constraints.

I am aware of the LSEI fortran routine, which is accessible from R, python, and other software. the solve.QP from R quadprog package is also perfect for this task. Finaly, the SLSQP fortran routine does also the job properly.

However i need arbitrary precision and for that i'm using the mpmath pyhton library. Are you aware of some implementaiton of one of these algorithmes for mpmath ?

lrnv
  • 1,038
  • 8
  • 19
  • I know of quad precision solvers but I am not aware of arbitrary precision solvers. I guess they would be prohibitively slow. There are rational MIP solvers. Maybe that can be used to solve the KKT conditions. – Erwin Kalvelagen Jul 01 '20 at 09:21
  • @ErwinKalvelagen Yeah i did not found anything either. I agree, they will probably be very very slow, but so usefull :) – lrnv Jul 02 '20 at 12:02

0 Answers0