Originally posted by: SystemAdmin

[UDOPS said:]

The OPL documentation mentions the existence of a subset function, but I cannot find the syntax for it anywhere. Anyone know how to test the validity of a subset, say for example in an assert statement?
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

[afleischer said:]

Hi,

there is no direct method in OPL 6.0 but there are many ways to check this.

For example to check that set A is a subset of B

you could write

assert card(A inter B)==card(A)

or

assert forall(i in A) i in B




#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

[UDOPS said:]

Thanks. I like the latter method better, because the first method leaves some doubt in my mind about elements counts versus actual membership checks. Is there a way to test equivalence of sets? Could I have "assert {A inter B} == A" ?
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

[afleischer said:]

equivalence could be checked thru both subset

assert forall(i in A) i in B
assert forall(i in B) i in A

#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

[UDOPS said:]

OK, lets make this harder. Neither of these syntaxes work. Any suggestions?

assert forall(r in TrainData) pathvalid: forall(i in minl(r.origin, r.destination)..maxl(r.origin, r.destination)) <i, c, t, n> in Profile;

assert forall(r in TrainData) pathvalid: forall(i in minl(r.origin, r.destination)..maxl(r.origin, r.destination)) i in {<p.block> | p in Profile};


//Profile is a 4 element tuple set.
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

[afleischer said:]

could

assert forall(r in TrainData) pathvalid: forall(i in minl(r.origin, r.destination)..maxl(r.origin, r.destination))
card({<i, c, t, n> | <i, c, t, n> in Profile})!=0;

work better?

Alex
#DecisionOptimization
#OPLusingCPLEXOptimizer

Originally posted by: SystemAdmin

[UDOPS said:]

Thank you, that worked.

There seems to be a pattern here. I have to start considering aggregate functions more often.
#DecisionOptimization
#OPLusingCPLEXOptimizer