odow
commented
3 years ago @andreaslundell SHOT is working from Julia, but there are a few bugs to sort out.
LogLevel is the opposite convention to every other solver
I was about to open an issue that set_optimizer_attribute(model, "Output.Console.LogLevel", 0) was being ignored, when I realized that =6 is "off" and =0 is "trace". Why this way around?
SHOT has not been compiled with support for selected MIP solver.
See the log below. I thought we compiled it for Cbc? And it finds Cbc, so why the warning?
sol file not written for infeasible problems
SHOT doesn't write a sol file for infeasible problems. How can we detect that a problem is infeasible instead of the solver aborting due to an error?
It would be nicer if it wrote a sol file with objno 0 200.
julia> using JuMP, AmplNLWriter, SHOT_jll
julia> model = Model(bridge_constraints = false) do
AmplNLWriter.Optimizer(SHOT_jll.amplexe)
end
A JuMP Model
Feasibility problem with:
Variables: 0
Model mode: AUTOMATIC
CachingOptimizer state: EMPTY_OPTIMIZER
Solver name: AmplNLWriter
julia> set_optimizer_attribute(model, "Output.Console.LogLevel", 0)
0
julia> @variable(model, x >= 0.1, Int)
x
julia> @constraint(model, x <= 0)
x ≤ 0.0
julia> optimize!(model)
╶ Supporting Hyperplane Optimization Toolkit (SHOT) ──────────────────────────────────────────────────────────────────╴
Andreas Lundell and Jan Kronqvist, Åbo Akademi University, Finland.
See documentation for full list of contributors and utilized software libraries.
Version: 1.0.1. Git hash: d2c99ba4-dirty. Released: Jul 7 2021.
For more information visit https://shotsolver.dev
╶ Modeling system ────────────────────────────────────────────────────────────────────────────────────────────────────╴
Modeling system: AMPL
Problem read from file: /var/folders/bg/dzq_hhvx1dxgy6gb5510pxj80000gn/T/jl_jptRUf/model.nl
SHOT has not been compiled with support for selected MIP solver.
Cbc selected as MIP solver.
Creating dual problem
Dual problem created
╶ Problem instance ───────────────────────────────────────────────────────────────────────────────────────────────────╴
Original Reformulated
Problem classification: MILP, convex MILP, convex
Objective function direction: minimize minimize
Objective function type: nonlinear linear
Number of constraints: 1 1
- linear: 1 1
Number of variables: 1 1
- integer: 1 1
╶ Options ────────────────────────────────────────────────────────────────────────────────────────────────────────────╴
No options file specified.
Options specified:
- Dual.MIP.Solver = 2
- Dual.TreeStrategy = 0
- Model.Reformulation.Quadratics.Strategy = 0
- Output.Console.LogLevel = 0
- Output.OutputDirectory = 0
- Primal.Tolerance.TrustLinearConstraintValues = false
Dual strategy: Multi-tree
- cut algorithm: ESH
- solver: Cbc 2.10.5
Primal NLP solver: Ipopt 3.13.4 (with default linear solver)
Activating LP strategy
╶ Main iteration step ────────────────────────────────────────────────────────────────────────────────────────────────╴
Iteration │ Time │ Dual cuts │ Objective value │ Objective gap │ Current solution
#: type │ tot. │ + | tot. │ dual | primal │ abs. | rel. │ obj.fn. | max.err.
╶─────────────────┴────────┴─────────────┴─────────────────────────┴───────────────────┴──────────────────────────────╴
Solving dual problem.
Dual problem solved with return code: 2
Dual solver reports no solutions found.
1: LP-I 0.01 -inf. | inf. inf. | inf. | inf.
╶ Solution report ────────────────────────────────────────────────────────────────────────────────────────────────────╴
Terminated since the dual problem is infeasible.
Problem is infeasible.
Objective bound (minimization) [dual, primal]: [-1.79769e+308, 1.79769e+308].
Objective gap absolute / relative: inf / inf.
Fulfilled termination criteria:
Unfulfilled termination criteria:
- absolute objective gap tolerance inf > 0.001
- relative objective gap tolerance inf > 0.001
- maximal constraint tolerance 1.79769e+308 > 1e-08
- iteration limit 1 <= 200000
- solution time limit (s) 0.00771078 <= 1.79769e+308
Dual problems solved in main step: 1
- LP problems 1
Total solution time: 0.00773052
- problem reformulation: 4.7759e-05
- bound tightening: 1.0099e-05
- dual strategy: 0.0033561
- solving relaxed problems: 0.00154916
- root search for constraint cuts: 3.258e-06
- primal strategy: 0.000973026
- solving NLP problems: 0.000858522
╶─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╴
Results written to: /var/folders/bg/dzq_hhvx1dxgy6gb5510pxj80000gn/T/jl_jptRUf/model.osrl
julia> termination_status(model)
OTHER_ERROR::TerminationStatusCode = 24
andreaslundell
Let's see the damage.