Improve choosing dividing-axis

src/genz-malik.jl

@@ -114,16 +114,17 @@ error `E` (via the given `norm`), and the suggested coordinate `k` ∈ `1:n`
 to subdivide next.
 """
 function (g::GenzMalik{n,T})(f, a::SVector{n}, b::SVector{n}, norm=vecnorm) where {n,T}
-    c = (a+b)*T(0.5)
-    Δ = (b-a)*T(0.5)
-    V = abs(prod(Δ))
+    c = T(0.5).*(a.+b)
+    Δ = T(0.5).*abs.(b.-a)
+    V = prod(Δ)
 
     f₁ = f(c)
 
     f₂ = zero(f₁)
     f₃ = zero(f₁)
     kdivide = 1
     maxdivdiff = zero(norm(f₁))
+    ε = eps(typeof(maxdivdiff))
     twelvef₁ = 12f₁
     for i = 1:n
         p₂ = Δ .* g.p[1][i]
@@ -135,9 +136,11 @@ function (g::GenzMalik{n,T})(f, a::SVector{n}, b::SVector{n}, norm=vecnorm) wher
         # fourth divided difference: f₃ᵢ-2f₁ - 7*(f₂ᵢ-2f₁),
         # where 7 = (λ₃/λ₂)^2 [see van Dooren and de Ridder]
         divdiff = norm(f₃ᵢ + twelvef₁ - 7*f₂ᵢ)
-        if divdiff > maxdivdiff
+        if (δ = divdiff - maxdivdiff) > ε
             kdivide = i
             maxdivdiff = divdiff
+        elseif abs(δ) < ε && Δ[i] > Δ[kdivide]
+            kdivide = i
         end
     end