Originally posted by: derekxu11

The decision variable in my model is constant, which is used as a float numbers in Python

But due to the accuracy, the result always come out like follows

t121 = 5.799900
t122 = 5.000000
t123 = 6.000000
t124 = 0.000100

Is there any way to solve this problem? Such as control the calculating accuracy or floating numbers accuracy?

It is unclear which of the numbers you posted is considered problematic.

In any case, parameters that control numerical accuracy in CPLEX are emphasis.numerical, simplex.tolerances.feasibility, and simplex.tolerances.optimality. I am not optimistic they are going to change a lot, though, since your numbers do not look that bad. What numbers what you like to see instead in the above cases?