How to demonstrate the calculation process and progress of the solver LINDOGlobal.

The calculation process for most problems is time-consuming and the calculation progress cannot be mastered.

How to demonstrate the calculation process and progress of the solver LINDOGlobal.

The calculation process for most problems is time-consuming and the calculation progress cannot be mastered.

To see a progress log from the LINDO Global solver while it is running, set the option `outlev=1`

. If you are not currently setting any options for LINDO Global, then you can specify the option like this:

```
option solver lindoglobal;
option lindoglobal_options 'outlev=1';
```

(If you are already setting a `lindoglobal_options`

string, then add `outlev=1`

to it.) For more information, visit our LINDO Global for AMPL page.

Thank you very much, but I still have no expectation of the solution time. May I ask which label I should focus on, and how to improve the speed of solving nonlinear problems.

In LINDO Global’s `outlev=1`

output, there is some information on how much time the solver has taken so far. There is no information on how long the total solution time will be, however, because there is no way for the solver’s algorithm to know that.

Nonlinear global optimization is a very hard problem, so there can be no simple advice for improving the speed of the solver. If you can post a file containing all of the solver’s output, it may be possible to give some general advice. However, you may need to stop the solver before it has proved optimality, and take the best solution it has found – which may be an optimal solution, or very close to one. You can set LINDO Global’s options as, for example,

```
option lindoglobal_options 'outlev=1 maxtime=1000';
```

to specify a maximum run time in seconds, after which the best solution so far will be returned.

You could also experiment with other solvers. To get advice on other solvers, you’ll need to say whether your model has integer-valued variables and whether it uses only smooth functions (or, post your model file).