Procedures and functions
The module mmnl overloads certain mathematical functions of the Mosel language, replacing an argument of type real by the types linctr and nlctr. The return value of these functions is of type nlctr. This means they can be used as operators in the definition of nonlinear constraints as shown in the example of Section Example: using mmnl for QCQP. The relevant functions are:
- Arithmetic functions:
- abs
- absolute value
- ceil
- rounding to the next largest integer
- exp
- natural exponent of the argument
- floor
- rounding to the next smallest integer
- ln
- natural logarithm of the argument
- log
- base 10 logarithm of the argument
- round
- rounding to the nearset integer
- sqrt
- positive square root of the argument
- sign
- sign of an expression (-1 if negative, 1 if positive, 0 othewise)
- fmin
- Minimum value of a list of expressions (this function accepts a variable number of arguments or a list)
- fmax
- Maximum value of a list of expressions (this function accepts a variable number of arguments or a list)
- Trigonometric functions:
Since these mathematical operators are fairly self-explanatory, we shall forego any more detailed documentation of these functions.
The following list gives an overview of all other functions and procedures defined by mmnl for which we give detailed descriptions later.
| clearinitvals |
Delete all initial value definitions.
|
| copysoltoinit |
Copy solution values to initial values.
|
| getsol |
Get the solution value of a nonlinear constraint.
|
| gettype |
Get the type of a nonlinear constraint.
|
| ishidden |
Test whether a constraint is hidden.
|
| pwlin |
Generate a piecewise linear function.
|
| sethidden |
Hide or unhide a nonlinear constraint.
|
| setinitval |
Set an initial value (start value) for a variable.
|
| setname |
Associate a matrix name to a nonlinear constraint.
|
| settype |
Set the type of a nonlinear constraint.
|