Overview of Operator Efficiency

Operator efficiency

In this section you will find an overview of the efficiency of all unary, binary and iterative operators in AIMMS.

Unary operators and functions

The unary operators and functions presented in this table are divided in two groups: sparse and dense.

• sparse: Here, when the argument is 0.0, the result is 0.0. The result needs to be computed only for those tuples for which the argument has a non-zero value.

• dense: Here, when the argument is 0.0, the result is not equal to 0.0. The results of all possible tuples need to be computed.

Table 25 Sparsen and dense unary operators and functions

sparse		dense
-	Sinh	NOT	Cos
Sin	Tanh	Cos	Cosh
Tan	ArcSin	Exp	ArcCos
Round	ArcTan	Log	ArcCosh
Floor	ArcSinh	Log10	Factorial
Ceil	ArcTanh
Trunc	Sqr
Sqrt

Binary operators

The binary operators presented in this table can be divided in three groups:

• intersection sparse: Here, when either of the arguments is 0.0, the result is 0.0. The result of only those tuples need to be computed where both arguments are not equal to 0.0. This corresponds to taking the intersection of the set of tuples for which the arguments are defined.

• union sparse: Here, when both arguments are 0.0, the result is 0.0. The result of only those tuples need to be computed where at least one of the arguments is not equal to 0.0. This corresponds to taking the union of the set of tuples for which the arguments are defined.

• dense: Here, when both arguments are 0.0, the result is not equal to 0.0. In this case, the expression needs to be evaluated for all possible combinations of values of the indices, unless these combinations are limited by a sparse operator elsewhere in the same expression. This corresponds to taking the Cartesian product of the ranges of the indices.

Table 26 Sparseness of binary operators

intersection	union	dense
*	+	^
$	-	/
ONLYIF	<>	=
AND	< > OR XOR	<= >= Permutation Combination

Iterative operators

The iterative operators presented in this table are divided in three groups as follows:

• sparse A value 0.0 of an argument does not influence the result and can safely be ignored. The iterative operator only considers existing entries of its argument.

• almost sparse A second 0.0 in the argument does not influence the result. The execution starts in a dense meaning that the iterative operator considers all possible tuples. However, after a first 0.0 has been encountered, execution continues in a sparse manner.

• dense A value 0.0 in the argument influences the result. The iterative operator considers all possible combinations.

Table 27 Sparseness of iterative operators

sparse	almost sparse	dense
Sum	Max	Mean	SampleDeviation
Prod	Min	GeometricMean	PopulationDeviation
Exists	ArgMax	HarmonicMean	Skewness
Forall	ArgMin	RootMeanSquare	Kurtosis
Count		Median	RankCorrelation