Benoît Legat
Benoît Legat
As discussed with @matbesancon during the JuMP-dev call, we could add a `PolyJuMP.Geometric.Optimizer` meta-solver that supports `ScalarPolynomialFunction` as input and rewrites into exponential conic form using geometric programming (and errors...
These should accept a `result` keyword https://github.com/jump-dev/PolyJuMP.jl/blob/8eb58029a2181cbd1ba0338118f5ca5415b83a2b/src/variable.jl#L3-L17 like https://github.com/jump-dev/JuMP.jl/blob/6901e8f98bf24242a141e003e42530fa90f33c3c/src/aff_expr.jl#L493-L495
Trying to make this work `operate!(add_dot, BigInt(), [A, A], [A, A])` where `A = ones(BigInt, 2, 2)` to remove the need for such hacks: https://github.com/kalmarek/PropertyT.jl/blob/master/src/sos_sdps.jl#L144-L173 cc @kalmarek
It is used https://github.com/JuliaAlgebra/MultivariatePolynomials.jl/blob/c9cbca9fcc0d9d6bafced0d983d20e673a50a1bd/src/division.jl#L271 and it should be used https://github.com/JuliaAlgebra/MultivariatePolynomials.jl/blob/c9cbca9fcc0d9d6bafced0d983d20e673a50a1bd/src/division.jl#L53 Once it's done, we should see nice improvements for https://github.com/JuliaAlgebra/MultivariatePolynomials.jl/issues/194
Some implementations in Base/LinearAlgebra/SparseArrays have the following assumption: 1. `zero(T) isa T` 2. `one(T) isa T` 3. `-(::T) isa T` 4. `+(::T, ::T) isa T` 5. `*(::T, ::T) isa T`...
e.g. `operate_to!!(::Int, ::typeof(//), ::Int, ::Int)` doesn't work cc @kalmarek
It requires https://github.com/jump-dev/MathOptInterface.jl/pull/1046 which will be included in MOI v0.9.15
This also fixes a bug: the `result_index` were ignored and we would invariably query the first solution. - [ ] Test for starting values - [ ] Test for the...
Came up in https://github.com/jump-dev/SumOfSquares.jl/pull/283
It is currently quite difficult to use CSDP from another language as there is no API to play with the structures. In Julia, we mirrored the C-structures and modify them...