[Bridges] add ComplementsToScalarNonlinearFunctionBridge

[odow]

Closes #2589

This would be a replacement for Complementarity.jl (cc @chkwon). It's not a perfect replacement because there's no flexibility in how we reformulate, but it's good enough for basic stuff:

julia> using JuMP, Ipopt

julia> begin
           model = Model(Ipopt.Optimizer)
           set_silent(model)
           @variable(model, x)
           @variable(model, z)
           @variable(model, w >= 0)
           @objective(model, Min, (x - 3.5)^2 + (z + 4)^2)
           @constraint(model, z - 3 + 2*z*w == 0)
           @constraint(model, -(z^2 - x) ⟂  w)
           optimize!(model)
           solution_summary(model)
       end
* Solver : Ipopt

* Status
  Result count       : 1
  Termination status : LOCALLY_SOLVED
  Message from the solver:
  "Solve_Succeeded"

* Candidate solution (result #1)
  Primal status      : FEASIBLE_POINT
  Dual status        : FEASIBLE_POINT
  Objective value    : 2.82500e+01

* Work counters
  Solve time (sec)   : 8.78320e-02
  Barrier iterations : 303

Current reformulation is really just a proof of concept to check things work. Numerically, we might be better off doing this reformulation though. It just required more lines of code so I didn't do it to begin with.

y - F == 0  # G-in-EqualTo
if isfinite(l)
    (x - l) * y <= 0.0  # ScalarQuadratic-in-LessThan
else
    y <= 0.0  # VariableIndex-in-LessThan
end
if isfinite(u)
    (x - u) * y <= 0.0  # ScalarQuadratic-in-LessThan
else
    y >= 0.0.  # VariableIndex-in-GreaterThan
end

[odow]

Add as a package, not a bridge