Skip to main content

Numerical trouble encountered when solving convex QCQP



1 comment

  • David Torres Sanchez
    Gurobi Staff Gurobi Staff

    Hi Braun,

    Your objective function range increases with every iteration until it gets too large to be manageable. At the last iteration, we get:

    Coefficient statistics:
      Matrix range     [1e+00, 3e+03]
      QMatrix range    [3e-02, 1e-01]
      QLMatrix range   [1e-01, 1e+00]
      Objective range  [2e-14, 1e+03]
      QObjective range [1e-02, 1e-02]
      Bounds range     [1e+00, 1e+00]
      RHS range        [5e-01, 5e-01]
      QRHS range       [2e-02, 2e-02]

    Changing lambda to 10 also works.
    You need to try and scale these coefficients to reduce this range.
    Please see: Guidelines for Numerical Issues
    As suggested in that guide, using Presolve = 0 also works for this model.



Please sign in to leave a comment.