Failure with CachingOptimizer #2384
Comments
|
Here's a reproducible example: julia> import MathOptInterface as MOI
julia> import CSDP
julia> inner = MOI.instantiate(CSDP.Optimizer; with_bridge_type = Float64)
MOIB.LazyBridgeOptimizer{MOIU.CachingOptimizer{CSDP.Optimizer, MOIU.UniversalFallback{MOIU.Model{Float64}}}}
with 0 variable bridges
with 0 constraint bridges
with 0 objective bridges
with inner model MOIU.CachingOptimizer{CSDP.Optimizer, MOIU.UniversalFallback{MOIU.Model{Float64}}}
in state EMPTY_OPTIMIZER
in mode AUTOMATIC
with model cache MOIU.UniversalFallback{MOIU.Model{Float64}}
fallback for MOIU.Model{Float64}
with optimizer CSDP.Optimizer
julia> MOI.set(inner, MOI.Silent(), true)
julia> model = MOI.Utilities.CachingOptimizer(MOI.Utilities.Model{Float64}(), inner)
MOIU.CachingOptimizer{MOIB.LazyBridgeOptimizer{MOIU.CachingOptimizer{CSDP.Optimizer, MOIU.UniversalFallback{MOIU.Model{Float64}}}}, MOIU.Model{Float64}}
in state EMPTY_OPTIMIZER
in mode AUTOMATIC
with model cache MOIU.Model{Float64}
with optimizer MOIB.LazyBridgeOptimizer{MOIU.CachingOptimizer{CSDP.Optimizer, MOIU.UniversalFallback{MOIU.Model{Float64}}}}
with 0 variable bridges
with 0 constraint bridges
with 0 objective bridges
with inner model MOIU.CachingOptimizer{CSDP.Optimizer, MOIU.UniversalFallback{MOIU.Model{Float64}}}
in state EMPTY_OPTIMIZER
in mode AUTOMATIC
with model cache MOIU.UniversalFallback{MOIU.Model{Float64}}
with 1 optimizer attribute
fallback for MOIU.Model{Float64}
with optimizer CSDP.Optimizer
julia> x = MOI.add_variables(model, 2)
2-element Vector{MathOptInterface.VariableIndex}:
MOI.VariableIndex(1)
MOI.VariableIndex(2)
julia> MOI.set(model, MOI.ObjectiveSense(), MOI.MIN_SENSE)
julia> f = 1.0 * x[1] + 2.0 * x[2]
0.0 + 1.0 MOI.VariableIndex(1) + 2.0 MOI.VariableIndex(2)
julia> MOI.set(model, MOI.ObjectiveFunction{typeof(f)}(), f)
julia> MOI.add_constraint.(model, x, MOI.GreaterThan(1.0))
2-element Vector{MathOptInterface.ConstraintIndex{MathOptInterface.VariableIndex, MathOptInterface.GreaterThan{Float64}}}:
MathOptInterface.ConstraintIndex{MathOptInterface.VariableIndex, MathOptInterface.GreaterThan{Float64}}(1)
MathOptInterface.ConstraintIndex{MathOptInterface.VariableIndex, MathOptInterface.GreaterThan{Float64}}(2)
julia> MOI.optimize!(model)
julia> MOI.delete(model, x[1])
julia> MOI.optimize!(model)
julia> print(inner)
Minimize ScalarAffineFunction{Float64}:
1.0 + 2.0 v[-2]
Subject to:
VariableIndex-in-GreaterThan{Float64}
v[-2] >= 1.0
julia> @assert isapprox(MOI.get(model, MOI.ObjectiveValue()), 2.0)
ERROR: AssertionError: isapprox(MOI.get(model, MOI.ObjectiveValue()), 2.0)
Stacktrace:
[1] top-level scope
@ REPL[283]:1Note that the objective of |
|
The other issue is: julia> using Test, Revise
julia> import MathOptInterface as MOI
julia> import CSDP
julia> inner = MOI.instantiate(CSDP.Optimizer, with_bridge_type = Float64)
MOIB.LazyBridgeOptimizer{MOIU.CachingOptimizer{CSDP.Optimizer, MOIU.UniversalFallback{MOIU.Model{Float64}}}}
with 0 variable bridges
with 0 constraint bridges
with 0 objective bridges
with inner model MOIU.CachingOptimizer{CSDP.Optimizer, MOIU.UniversalFallback{MOIU.Model{Float64}}}
in state EMPTY_OPTIMIZER
in mode AUTOMATIC
with model cache MOIU.UniversalFallback{MOIU.Model{Float64}}
fallback for MOIU.Model{Float64}
with optimizer CSDP.Optimizer
julia> model = MOI.Utilities.CachingOptimizer(MOI.Utilities.Model{Float64}(), inner)
MOIU.CachingOptimizer{MOIB.LazyBridgeOptimizer{MOIU.CachingOptimizer{CSDP.Optimizer, MOIU.UniversalFallback{MOIU.Model{Float64}}}}, MOIU.Model{Float64}}
in state EMPTY_OPTIMIZER
in mode AUTOMATIC
with model cache MOIU.Model{Float64}
with optimizer MOIB.LazyBridgeOptimizer{MOIU.CachingOptimizer{CSDP.Optimizer, MOIU.UniversalFallback{MOIU.Model{Float64}}}}
with 0 variable bridges
with 0 constraint bridges
with 0 objective bridges
with inner model MOIU.CachingOptimizer{CSDP.Optimizer, MOIU.UniversalFallback{MOIU.Model{Float64}}}
in state EMPTY_OPTIMIZER
in mode AUTOMATIC
with model cache MOIU.UniversalFallback{MOIU.Model{Float64}}
fallback for MOIU.Model{Float64}
with optimizer CSDP.Optimizer
julia> x = MOI.add_variable(model)
MOI.VariableIndex(1)
julia> c1 = MOI.add_constraint(model, x, MOI.GreaterThan(1.0))
MathOptInterface.ConstraintIndex{MathOptInterface.VariableIndex, MathOptInterface.GreaterThan{Float64}}(1)
julia> c2 = MOI.add_constraint(model, x, MOI.LessThan(1.0))
MathOptInterface.ConstraintIndex{MathOptInterface.VariableIndex, MathOptInterface.LessThan{Float64}}(1)
julia> MOI.optimize!(model)
CSDP 6.2.0
This is a pure primal feasibility problem.
This is a pure dual feasibility problem.
Iter: 0 Ap: 0.00e+00 Pobj: 0.0000000e+00 Ad: 0.00e+00 Dobj: 0.0000000e+00
Iter: 1 Ap: 9.00e-01 Pobj: 0.0000000e+00 Ad: 1.00e+00 Dobj: 0.0000000e+00
Success: SDP solved
Primal objective value: 0.0000000e+00
Dual objective value: 0.0000000e+00
Relative primal infeasibility: 5.00e-11
Relative dual infeasibility: 0.00e+00
Real Relative Gap: 0.00e+00
XZ Relative Gap: 7.50e-10
DIMACS error measures: 5.00e-11 0.00e+00 0.00e+00 0.00e+00 0.00e+00 7.50e-10
julia> MOI.get(model, MOI.VariablePrimal(), x)
1.000000000025
julia> MOI.get(model, MOI.ConstraintPrimal(), c1)
1.000000000025
julia> MOI.get(model, MOI.ConstraintPrimal(), c2)
-2.5e-11The last value is |
|
Simplified to one julia> import MathOptInterface as MOI
julia> import CSDP
julia> cache = MOI.Utilities.Model{Float64}();
julia> x, c1 = MOI.add_constrained_variable(cache, MOI.GreaterThan(1.0));
julia> c2 = MOI.add_constraint(cache, 1.0x, MOI.EqualTo(2.0));
julia> model = MOI.instantiate(CSDP.Optimizer, with_bridge_type = Float64);
julia> index_map, _ = MOI.optimize!(model, cache);
CSDP 6.2.0
This is a pure primal feasibility problem.
Iter: 0 Ap: 0.00e+00 Pobj: 0.0000000e+00 Ad: 0.00e+00 Dobj: 0.0000000e+00
Iter: 1 Ap: 1.00e+00 Pobj: 0.0000000e+00 Ad: 1.00e+00 Dobj: 9.0000000e+00
Success: SDP solved
Primal objective value: 0.0000000e+00
Dual objective value: 0.0000000e+00
Relative primal infeasibility: 0.00e+00
Relative dual infeasibility: 5.00e-11
Real Relative Gap: 0.00e+00
XZ Relative Gap: 5.00e-11
DIMACS error measures: 0.00e+00 0.00e+00 5.00e-11 0.00e+00 0.00e+00 5.00e-11
julia> MOI.get(model, MOI.VariablePrimal(), index_map[x])
2.0
julia> MOI.get(model, MOI.ConstraintPrimal(), index_map[c1])
2.0
julia> MOI.get(model, MOI.ConstraintPrimal(), index_map[c2])
1.0
julia> MOI.Bridges.print_active_bridges(model)
* Supported objective: MOI.ScalarAffineFunction{Float64}
* Supported constraint: MOI.ScalarAffineFunction{Float64}-in-MOI.EqualTo{Float64}
* Unsupported variable: MOI.GreaterThan{Float64}
| bridged by:
| MOIB.Variable.VectorizeBridge{Float64, MOI.Nonnegatives}
| may introduce:
| * Supported variable: MOI.Nonnegatives |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
This seems like a bug in
CachingOptimizerorLazyBridgeOptimizer. CSDP passes the tests without this additional cache and when adding this cache (which gives the setting of the JuMP backend, we have a failure). It seems SDPA has the same issue since it's failing on the same tests with the same wrong answers6and0.produces
The text was updated successfully, but these errors were encountered: