Minimization with constraint involving logs in Julia

回答済み

Albert John Crosby

2024年03月13日 21:46 編集済み

• 
• 
• 

I want to solve \(\frac{1}{N} \sum_{i=1}^N x_i^2\) subject to \(\frac{1}{N} \sum_{i=1}^N \log x_i \geq 0\). 

In the above, N is the size of the vector x whose \(i^{th}\) element is \(x_i\). I think the solution should be \(x_j = 1\), for all \(j=1,\ldots, N\). How do I code this in Julia? I appreciate that I could replace the constraint involving a sum by a product but in the true application I have in mind (which involves logs) this may not be possible. Basically I want to minimize a function which is convex and straightforward but with a constraint involving logs. I tried the following but it doesn't work. In the below, ncol is the size of n (which is denoted x above)

model = Model(Gurobi.Optimizer)
@variable(model, 1000.0 >= n[1:ncol] >= 0.001)
@objective(model, Min, sum(n[i]^2 for i in eachindex(n)) / ncol)
@expression(model, expr, sum(log(n[i]) for i in 1:ncol))
@NLconstraint(model, expr >= 0.0)
optimize!(model)

 

  Can anyone tell me how I could incorporate a constraint on the log of x? The code is Julia.

Thank you. 

 

関連

• Nonlinear
• Julia

0

• 

  David Torres Sanchez

  • Gurobi Staff

  2024年03月18日 11:56 編集済み

  See a similar post in the Julia discourse. In your case, you can use the GRBaddgenconstrLog which is documented here: GRBaddgenconstrLog - C API.

  Cheers, 
  David

  0

サインインしてコメントを残してください。