Originally posted by: OR_Nicole

Hi, I am relatively new to using CPLEX with Java API.  I am writing a larger multi-objective program but am having a fundamental problem.  I am defining my x variables as float type values between 0 and 1, however, the optimization results give me values such as 1.0000005883164844.  How do I prevent this from happening?  I defined my x variables as follows:

Thanks for any help!

This is not actually a problem. Due to the fact that floating point arithmetic is inexact, CPLEX tolerates small violations of constraints (including bounds) and small violations of integrality (by integer variables). The default constraint tolerance is 1e-6, and your value of 1.000000588... exceeds the upper bound by less than that.

I would suggest just rounding that value to 1.0. If you really want to reduce bound violations, you can change the value of the parameter IloCplex.Param.Simplex.Tolerances.Feasibility to something smaller than 1e-6 (by invoking the setParam() method on your IloCplex object. If you set it too small, though, CPLEX might possibly declare your problem infeasible when it is actually feasible.

Originally posted by: OR_Nicole

Thank you, Professor Rubin.  Indeed, I adjusted the feasibility tolerances and it helped reduce the bound violations. By the way, your blog is a great resource! 

You're welcome, and thank you (in that order).