I have a mixed-integer optimization problem with two variables: A binary integer, x, and a float, P_G I was a little surprised that the value matrix of the binary integer variable was of type float64 and not int64: julia> value.(x) 10×100 Array{Float64,2}: 0.0 1.0 1.0 0.0 0.0 -0.0 0.0 … …

Binary integer variable returned as float64

odow November 23, 2020, 8:02pm 2

This is expected behavior. All variables are Float64 and solvers use tolerances for checking integrality. Read more:

4 Likes

Topic		Replies	Views
Type of binary variables in constraints Optimization (Mathematical)	4	1612	June 13, 2019
Float64 and binary variables in constraints Optimization (Mathematical)	2	580	June 27, 2018
(Jump, Gurobi)Binary Variables end up with values no equal 1 or 0 Optimization (Mathematical)	8	1420	August 5, 2021
Convert ILP variables to strict integer values in JuMP Optimization (Mathematical) jump	1	381	November 23, 2021
Ipopt solver: no method matching setvartype! Optimization (Mathematical)	1	827	May 23, 2018