16.10.2013 Views

GAMS/PATH User Guide Version 4.3

GAMS/PATH User Guide Version 4.3

GAMS/PATH User Guide Version 4.3

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

of preprocessing and can be used to solve constrained systems of nonlinear<br />

equations. The output for <strong>PATH</strong>C is identical to the main distribution described<br />

in Section 2.1 with additional output for preprocessing. The options<br />

are the same between the two versions.<br />

2.5.1 Preprocessing<br />

The preprocessor is work in progress. The exact output in the final<br />

version may differ from that given below.<br />

The purpose of a preprocessor is to reduce the size and complexity of<br />

a model to achieve improved performance by the main algorithm. Another<br />

benefit of the analysis performed is the detection of some provably unsolvable<br />

problems. A comprehensive preprocessor has been incorporated into <strong>PATH</strong>C<br />

as developed in [18].<br />

The preprocessor reports its finding with some additional output to the<br />

log file. This output occurs before the initial point statistics. An example of<br />

the preprocessing on the forcebsm model is presented below.<br />

Zero: 0 Single: 112 Double: 0 Forced: 0<br />

Preprocessed size: 72<br />

The preprocessor looks for special polyhedral structure and eliminates variables<br />

using this structure. These are indicated with the above line of text.<br />

Other special structure is also detected and reported.<br />

On exit from the algorithm, we must generate a solution for the original<br />

problem. This is done during the postsolve. Following the postsolve, the<br />

residual using the original model is reported.<br />

Postsolved residual: 1.0518e-10<br />

This number should be approximately the same as the final residual reported<br />

on the presolved model.<br />

2.5.2 Constrained Nonlinear Systems<br />

Modelers typically add bounds to their variables when attempting to solve<br />

nonlinear problems in order to restrict the domain of interest. For example,<br />

many square nonlinear systems are formulated as<br />

F (z) =0,ℓ≤ z ≤ u,<br />

33

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

Saved successfully!

Ooh no, something went wrong!