MethodError: no method matching isless(::Int64, ::AffExpr)

Specific Domains Optimization (Mathematical)
,
1

I’m receiving an error MethodError: no method matching isless(::Int64, ::AffExpr) while trying to run my code in the constraints “@constraint(model, con8[k in 1:n,i in 1:m; i==m], T_k[k]==max(C_ki[k,i]-U_k[k],0))”

2

Please see this link, I believe that you have to register the function max before using it in a model.

1 Like
3

@vinicius_de_lima is correct. You cannot use arbitrary Julia functions in @constraint. You must formulate a mixed-integer linear program.

Instead of

model = Model()
@variable(model, x)
@variable(model, t)
@constraint(model, t == max(x, 0))

do

model = Model()
@variable(model, x)
@variable(model, t >= 0)
@constraint(model, t >= x)

or

model = Model()
@variable(model, x)
@variable(model, t)
register(model, :max, 2, max; autodiff = true)
@NLconstraint(model, t == max(x, 0))
4

Thank you so much for the reply. I will follow your lead and let you know if this is working or not.

1 Like
5

Thank you for your help, it worked. Now, I’m facing one more problem in which there are n non-negative numbers (i.e., tn>=0) and I want to find out how many numbers out of n are positive (i.e., tn>0). Please tell me how to calculate these numbers in JULIA as an expression in my MILP model.

6

The mosek modeling cookbook covers this:
https://docs.mosek.com/modeling-cookbook/mio.html#implication-of-positivity

If you have trouble, start a new post with code of what you’ve tried.

7

Thanks for the help.