Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

DE rastrigin tests sometimes randomly fail #97

Open
ianatol opened this issue Mar 18, 2022 · 0 comments
Open

DE rastrigin tests sometimes randomly fail #97

ianatol opened this issue Mar 18, 2022 · 0 comments

Comments

@ianatol
Copy link

ianatol commented Mar 18, 2022

The example below fails ~ 1 in every 10 times on Julia master, with the following error:

using StableRNGs
using Random
using Test
using Evolutionary

rng = StableRNG(42)

N = 3
P = 100
initState = ()->rand(rng, N)

function test_result(result::Evolutionary.EvolutionaryOptimizationResults, N::Int, tol::Float64)
    fitness = minimum(result)
    extremum = Evolutionary.minimizer(result)
    if round(fitness) == 0
        @test extremum  zeros(N) atol=tol
        @test fitness  0.0 atol=tol
    else
        # @warn("Found local minimum!!!")
        @test sum(abs, extremum) < N
    end
end

# Objective function
function rastrigin(x::AbstractVector{T}) where {T <: AbstractFloat}
    n = length(x)
    return 10n + sum([ x[i]^2 - 10cos(convert(T,2π*x[i])) for i in 1:n ])
end

selections = [:rand=>random, :perm=>permutation, :rndoff=>randomoffset, :best=>best]
mutations = [:exp=>EXPX(0.5), :bin=>BINX(0.5)]
opts = Evolutionary.Options(rng=rng, successive_f_tol=25)
@testset "DE settings" for (sn,ss) in selections, (mn,ms) in mutations, n in 1:2
    Random.seed!(rng, 1)
    result = Evolutionary.optimize( rastrigin, initState,
        DE(
            populationSize = P,
            n=n,
            selection = ss,
            recombination = ms,
            F = 0.9
           ),
        opts
    )
    println("DE/$sn/$n/$mn(F=0.9,Cr=0.5) => F: $(minimum(result)), C: $(Evolutionary.iterations(result))")
    #println(Evolutionary.minimizer(result))
    test_result(result, N, 1e-4)
end

Error:

Test Summary: | Pass  Total  Time
DE settings   |    2      2  0.4s
DE/rndoff/2/exp(F=0.9,Cr=0.5) => F: 0.00038250130044303887, C: 61
DE settings: Test Failed at REPL[9]:5
  Expression: ≈(extremum, zeros(N), atol = tol)
   Evaluated: [-0.0008353205026887522, 0.0007904746949007734, 0.0026797512063458733] ≈ [0.0, 0.0, 0.0] (atol=0.0001)
Stacktrace:
 [1] macro expansion
   @ ~/master/usr/share/julia/stdlib/v1.9/Test/src/Test.jl:464 [inlined]
 [2] test_result(result::Evolutionary.EvolutionaryOptimizationResults{DE{typeof(randomoffset), Evolutionary.var"#expxvr#152"{Evolutionary.var"#expxvr#151#153"{Float64}}}, Vector{Float64}, Float64}, N::Int64, tol::Float64)
   @ Main ./REPL[9]:5
DE settings: Test Failed at REPL[9]:6
  Expression: ≈(fitness, 0.0, atol = tol)
   Evaluated: 0.00038250130044303887 ≈ 0.0 (atol=0.0001)
Stacktrace:
 [1] macro expansion
   @ ~/master/usr/share/julia/stdlib/v1.9/Test/src/Test.jl:464 [inlined]
 [2] test_result(result::Evolutionary.EvolutionaryOptimizationResults{DE{typeof(randomoffset), Evolutionary.var"#expxvr#152"{Evolutionary.var"#expxvr#151#153"{Float64}}}, Vector{Float64}, Float64}, N::Int64, tol::Float64)
   @ Main ./REPL[9]:6
Test Summary: | Fail  Total  Time
DE settings   |    2      2  0.5s
Test Summary: | Fail  Total  Time
DE settings   |    2      2  0.6s
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant