Make multiplication with a scalar fall back to scale

[schillic]

julia> X = rand(Interval)
Interval{Float64,IntervalArithmetic.Interval{Float64}}([0.419371, 1.82231])

julia> 2 * X
LinearMap{Float64,Interval{Float64,IntervalArithmetic.Interval{Float64}},Int64,SparseArrays.SparseMatrixCSC{Int64,Int64}}(
  [1, 1]  =  2, Interval{Float64,IntervalArithmetic.Interval{Float64}}([0.419371, 1.82231]))

I would have expected that I get a scaled interval here.

I propose to redefine *(α::Number, X::LazySet) = scale(α, X).

[mforets]

👍 why not. it is a special case. not sure about the general definition over LazySet though, because that needs to implement scale on every lazy set. on the other hand, the LinearMap only needs to have defined the support function so i think we should keep that behavior

[schillic]

Okay, maybe we can automatically define the methods for those sets that have scale defined.

[schillic]

I decided against that. This would make the library inconsistent.