[Ipopt] Question about warm start in combination with other IPOPT options

[Lemoine, Roberto]

Dear all,

one question. Assume that all the information about the degrees of freedom and Lagrange multipliers (e.g. from a former solution) is given to IPOPT (C++ interface) to do a warm start. Assume that the following options are always specified:

warm_start_init_point:  "yes"
mu_init:                (some value, e.g. 1.0e-3)

Now assume that one or all of the following additional options are also given:

expect_infeasible_problem:       "yes"
recalc_y:                        "yes"
(optionally: recalc_y_feas_tol:  (some value, e.g. 1.0e-2)

Do these additional options affect the behavior of the warm start? e.g. Can it happen that any information about degrees of freedom or multipliers, or any significant step(s) of the warm start algorithm is overridden?

Thank you very much.

Best regards,

Roberto Lemoine

-------------- next part --------------
A non-text attachment was scrubbed...
Name: winmail.dat
Type: application/ms-tnef
Size: 2938 bytes
Desc: not available
URL: <http://list.coin-or.org/pipermail/ipopt/attachments/20160913/6fc50899/attachment-0001.bin>