Hi,
I am solving an optimization problem of the following interface:
find max A such that
x^A + y < = 100 for all 0 <= x <= y <= 1
How can I write the following type of constraint in JuMP? And how it is better to define that for non-integer parameter A we set 0^A == 0?
Isn’t the answer A = \infty? 1^\infty + 1 = 2 \le 100.
Thank you for your reply,
yes , it’s better to change the 1 in constraint condition by 2, for example :).
In which case the answer is the solution to 2^A + 2 = 100 \implies A = \log_2 98.
I’ve wrote in the beginning that this problem is just similar to mine in interface. Of course, no need to use computers to solve directly the problem I’ve mentioned. I want to know the sintaxis of writing constraint in JuMP of the mentioned type.
You generally can’t write “for all” constraints in JuMP (that’s just not the way that solvers work).
You could write the inverse of the problem though. Find the minimum A such that the constraint is violated.
julia> using JuMP
julia> import Ipopt
julia> model = Model(Ipopt.Optimizer)
A JuMP Model
Feasibility problem with:
Variables: 0
Model mode: AUTOMATIC
CachingOptimizer state: EMPTY_OPTIMIZER
Solver name: Ipopt
julia> @variable(model, A >= 0.0)
A
julia> @variable(model, 0 <= x <= 2, start = 2)
x
julia> @variable(model, 0 <= y <= 2, start = 2)
y
julia> @NLconstraint(model, x^A + y >= 100)
(x ^ A + y) - 100.0 ≥ 0
julia> @objective(model, Min, A)
A
julia> optimize!(model)
This is Ipopt version 3.14.4, running with linear solver MUMPS 5.4.1.
Number of nonzeros in equality constraint Jacobian...: 0
Number of nonzeros in inequality constraint Jacobian.: 3
Number of nonzeros in Lagrangian Hessian.............: 3
Total number of variables............................: 3
variables with only lower bounds: 1
variables with lower and upper bounds: 2
variables with only upper bounds: 0
Total number of equality constraints.................: 0
Total number of inequality constraints...............: 1
inequality constraints with only lower bounds: 1
inequality constraints with lower and upper bounds: 0
inequality constraints with only upper bounds: 0
iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls
0 9.9999900e-03 9.70e+01 5.96e-01 -1.0 0.00e+00 - 0.00e+00 0.00e+00 0
1 1.6878344e-02 9.70e+01 1.04e+00 -1.0 5.56e+01 - 5.11e-04 3.53e-04h 1
2 1.4884628e+00 9.52e+01 2.37e+00 -1.0 1.40e+02 - 1.28e-04 1.05e-02h 1
3 7.6262848e+00 0.00e+00 2.76e+03 -1.0 4.91e+01 - 5.89e-02 1.25e-01h 4
4 7.6104472e+00 0.00e+00 1.59e+03 -1.0 1.23e-01 4.0 8.93e-02 1.00e+00h 1
5 7.6082893e+00 0.00e+00 2.89e+02 -1.0 3.65e-02 3.5 1.00e+00 1.00e+00h 1
6 7.6072385e+00 0.00e+00 5.54e+00 -1.0 4.09e-03 3.0 1.00e+00 1.00e+00h 1
7 7.6048327e+00 0.00e+00 3.52e+00 -1.0 9.51e-03 2.6 1.00e+00 1.00e+00f 1
8 6.4266817e+00 1.25e+01 3.79e+00 -1.0 8.60e+00 - 9.94e-01 1.37e-01f 1
9 6.7946110e+00 0.00e+00 1.03e+00 -1.0 2.82e+00 - 1.00e+00 1.00e+00f 1
iter objective inf_pr inf_du lg(mu) ||d|| lg(rg) alpha_du alpha_pr ls
10 6.8204764e+00 0.00e+00 2.01e-02 -1.0 3.81e+00 - 1.00e+00 1.00e+00h 1
11 6.6382699e+00 0.00e+00 1.87e-02 -2.5 6.93e+00 - 9.64e-01 9.59e-01f 1
12 6.6153426e+00 0.00e+00 4.30e-03 -3.8 8.16e-01 - 8.22e-01 1.00e+00h 1
13 6.6151594e+00 0.00e+00 7.80e-08 -3.8 8.22e-03 - 1.00e+00 1.00e+00h 1
14 6.6147160e+00 0.00e+00 1.30e-07 -5.7 1.01e-02 - 1.00e+00 1.00e+00h 1
15 6.6147098e+00 0.00e+00 2.80e-11 -8.6 1.74e-04 - 1.00e+00 1.00e+00h 1
Number of Iterations....: 15
(scaled) (unscaled)
Objective...............: 6.6147097558501411e+00 6.6147097558501411e+00
Dual infeasibility......: 2.8009594643663149e-11 2.8009594643663149e-11
Constraint violation....: 0.0000000000000000e+00 0.0000000000000000e+00
Variable bound violation: 1.9474300838595582e-08 1.9474300838595582e-08
Complementarity.........: 2.5931984733449229e-09 2.5931984733449229e-09
Overall NLP error.......: 2.5931984733449229e-09 2.5931984733449229e-09
Number of objective function evaluations = 18
Number of objective gradient evaluations = 16
Number of equality constraint evaluations = 0
Number of inequality constraint evaluations = 21
Number of equality constraint Jacobian evaluations = 0
Number of inequality constraint Jacobian evaluations = 16
Number of Lagrangian Hessian evaluations = 15
Total seconds in IPOPT = 0.009
EXIT: Optimal Solution Found.
julia> value(A)
6.614709755850141
julia> log2(98)
6.614709844115208