quasi
Posts:
10,226
Registered:
7/15/05


Re: disjunctive constraints in integer linear programming
Posted:
Apr 12, 2013 5:52 AM


Ray Vickson wrote: >If x is real valued (or even rational) the strict inequality >x < 0 cannot be modelled exactly; instead, one must replace it >by x <= =e, where e > 0 is "small". If x is also integer (<>0) >then, of course, x < 0 is the same as x <= 1, so that is >again OK.
Well, the OP did specify that it was an _integer_ linear program (actually two such programs the way I interpreted it), so based on that, the constraint x < 0 is just an alternate way of saying that x <= 1.
quasi

