CPXcheckax

Note

This is an advanced routine. Advanced routines typically demand a profound understanding of the algorithms used by ILOG CPLEX. Thus they incur a higher risk of incorrect behavior in your application, behavior that can be difficult to debug. Therefore, ILOG encourages you to consider carefully whether you can accomplish the same task by means of other Callable Library routines instead.

The routine CPXcheckax finds the L∞ norm of Ax - b. That is, this routine checks for numeric (roundoff) error in the computation of x (the resident solution) by putting it into that formula and determining which row has the maximum error from zero. This routine also returns, in one of its arguments, the index of the row with the maximum error from zero.

To get the L∞ norm for the scaled problem, set the parameter scalrimtype = 1.

The pointer to the ILOG CPLEX environment, as returned by CPXopenCPLEX.

A pointer to the CPLEX LP problem object, as returned by CPXcreateprob.

A pointer to the index of the row with the maximum absolute value in Ax - b. If no solution exists, *imax_p is set to -1.

An integer that indicates the type of scaling to be applied to the returned L∞ norm. When this parameter is equal to 0 (zero), the returned L∞ norm will be unscaled. Otherwise, the L∞ norm has the same scaling as that applied to the problem currently in memory.

If successful, the routine returns the L∞ norm of Ax - b, where x is the resident solution. If no such solution exists, -1.0 is returned.

CPXcheckax

Parameters:

Returns:

env	The pointer to the ILOG CPLEX environment, as returned by `CPXopenCPLEX`.
lp	A pointer to the CPLEX LP problem object, as returned by `CPXcreateprob`.
imax_p	A pointer to the index of the row with the maximum absolute value in Ax - b. If no solution exists, `*imax_p` is set to `-1`.
scalrimtype	An integer that indicates the type of scaling to be applied to the returned L∞ norm. When this parameter is equal to `0` (zero), the returned L∞ norm will be unscaled. Otherwise, the L∞ norm has the same scaling as that applied to the problem currently in memory.