Support for binary chromosomes?

[datawookie]

Hi,

I have been exploring Julia's GA capabilities, looking at GeneticAlgorithms and Evolutionary packages. I really enjoy the simple interface presented by Evolutionary and it also seems to be more advanced in terms of development. Great work and thank you for your efforts.

However, I am having trouble figuring out how I can implement a binary GA. I am using a knapsack problem as a simple illustrative example. I've written a fitness function which looks like this:

mass    = [1, 5, 3, 7, 2, 10, 5]
utility = [1, 3, 5, 2, 5,  8, 3]

function fitness(n::Vector{Int})
    total_mass = sum(mass .* n)
    return (total_mass <= 20) ? sum(utility .* n) : 0
end

but when I try to run the GA as follows

ga(fitness, 7)

I get

ERROR: wrong number of arguments
 in anonymous at /home/colliera/.julia/v0.3/Evolutionary/src/ga.jl:23
 in ga at /home/colliera/.julia/v0.3/Evolutionary/src/ga.jl:66

Looking at the debug output from ga() I see that the vectors being passed into my fitness function are floating point rather than binary (or at least a vector of integer 0 or 1). I'm not sure whether this is the source of the ERROR above, but it's certainly something that I need to resolve in order to get my problem up and running.

Is there support for binary chromosomes? Have I missed something?

Best regards,
Andrew.

[wildart]

It is just default selection and crossover functions are incorrect. Try some other functions. BTW, binary chromosomes supported as Vector{Bool}. I'll write some examples. Thanks for bringing this issue.

[wildart]

If you call your example as follows:

julia> ga(fitness, 7, selection=((x,n)->1:n))
([0.984655,0.684742,0.49106,0.204153,0.26621,0.660357,0.42944],13.804710257390415,700,0.0)

If you want binary chromosomes use Vector{Bool} to initialize the population (BitVector isn't supported}:

julia> ga(fitness, 7, selection=((x,n)->1:n), initPopulation = collect(randbool(7)))
(Bool[false,true,false,true,true,false,false],10,700,0)

[datawookie]

Thank you, that sorted out the problem perfectly!