Error on complex values in NLP

[odow]

Closes #1973

Related:

model = Model()
@variable(model, x >= 1im)
# ERROR: InexactError: Float64(0 + 1im)

Probably in lots of other places as well. What do we want to do about this?

[codecov]

Codecov Report

Merging #1978 into master will increase coverage by 0.02%.
The diff coverage is n/a.

@@            Coverage Diff            @@
##           master   #1978      +/-   ##
=========================================
+ Coverage   88.77%   88.8%   +0.02%     
=========================================
  Files          33      33              
  Lines        4260    4260              
=========================================
+ Hits         3782    3783       +1     
+ Misses        478     477       -1

Impacted Files	Coverage Δ
src/parse_nlp.jl	90.13% <ø> (+0.65%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 2201c60...4b6f121. Read the comment docs.

[mlubin]

Probably in lots of other places as well. What do we want to do about this?

Check all values that come into a macro, and throw an error if they match a condition that we know will throw an unhelpful internal error message (e.g., not being convertible to Float64). However, doing this well requires a bit of a refactoring of the macros so perhaps not high priority for now.

[Bondan000]

hi,
should this merge solve the problem with the complex numbers? i am using JuMP v0.20.1 and Julia v1.2.0 and still having this problem:
julia> mdl=Model()
julia> @variable(mdl, x >= 1*im)
ERROR: InexactError: Float64(0 + 1im)
is there any work around to use complex numbers while defining expressions?

[odow]

should this merge solve the problem with the complex numbers

No. JuMP does not support complex numbers.

is there any work around to use complex numbers while defining expressions?

Define two variables, one for the real part and one for the imaginary part

model = Model()
@variable(model, x_real >= 0)
@variable(model, x_imag >= 1)

See https://discourse.julialang.org/t/how-define-a-complex-jump-variable/14362

[Bondan000]

should this merge solve the problem with the complex numbers

No. JuMP does not support complex numbers.

is there any work around to use complex numbers while defining expressions?

Define two variables, one for the real part and one for the imaginary part
model = Model()
@variable(model, x_real >= 0)
@variable(model, x_imag >= 1)
See https://discourse.julialang.org/t/how-define-a-complex-jump-variable/14362

Thank you for the quick reply! will go through the discourse and try to solve my specific problem!