10.06.2016 Views

eldo_user

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

Code<br />

Table 13-3. Optimization Result Status Code (cont.)<br />

Message and meaning<br />

Optimization<br />

Final Diagnostic Conditions<br />

2 The current point cannot be improved. A sufficient decrease in the merit function<br />

could not be attained during the last line search. This may occur because the <strong>user</strong><br />

has requested an overly stringent accuracy.<br />

A sufficient decrease in the merit function could not be attained during the final<br />

line search. This sometimes occurs because an overly stringent accuracy has been<br />

requested, that is, TOL_OPT is too small. In this case verify the following to<br />

determine whether or not the final solution is acceptable:<br />

• The final values of Optimality are significantly small.<br />

• The final values of Feasibility are significantly small. This is set to zero if there<br />

are no equality or inequality constraints.<br />

An additional status code may be raised, of value: 0, 6, or 11.<br />

4 Optimization was not successful. The maximum number of simulation runs has<br />

been reached.<br />

The limiting number of simulations, determined by the optional parameter<br />

MAX_SIMUL has been reached.<br />

6 The current point is feasible and quite optimal, however, the required accuracy has<br />

not been achieved. The optimizer is within 1e-2 of satisfying the TOL_OPT (or<br />

TOL_GRAD) optimality tolerance.<br />

The optimizer is within 1x10 −2 of satisfying the TOL_OPT (or TOL_GRAD)<br />

optimality tolerance. For the next runs, check that TOL_OPT is not too small.<br />

7 The objective function may be unbounded on the feasible set. Alternatively, it may<br />

happen that the scaling of the problem is very poor.<br />

To avoid this situation, if the MINIMIZE or MAXIMIZE objectives are close to<br />

singularities, define bounds on some variables. Also use an appropriate scaling of<br />

the design objectives.<br />

8 The simulation process stopped the optimization. The optimization was<br />

abandoned.<br />

This message is raised only after an interrupt message (Ctrl+C).<br />

9 The current point cannot be improved. The problem is too hard to solve. Try with<br />

another starting point.<br />

This has to be considered as a failure of the optimizer regarding the treatment of<br />

constraint feasibilities. Please refer to “Normal and Elastic Modes of Termination”<br />

on page 659. Check that the final solution is not an acceptable point for the<br />

problem.<br />

10 The current point cannot be improved. The Hessian matrix has been reset too many<br />

times. It indicates that the line search failed to find a sufficiently better point for a<br />

large number of iterations.<br />

This message should appear very rarely. This is an algorithmic weakness of the<br />

method implemented in the Eldo optimizer SQP method. Try to start the<br />

optimization from a different point.<br />

Eldo® User's Manual, 15.3 657

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!