some utterly unimportant experiments

evo_test.jl

@@ -4,18 +4,22 @@
 
 load("evolib.jl")
 
-# define the objective function (in-place!)
-function rosenbrock(chr::Chromosome)
-    x1 = chr[1].gene
-    x2 = chr[2].gene
-    chr.fitness = 100*(x2-x1^2)^2+(1-x1)^2-50*((x1+1)^2+(x2-1)^2)-10*((x1-1.5)^2+(x2-2.5)^2)+exp(2*x2-5)
-end
+srand(123)
 
 # in case someone wants to calculate the objective function for a population
 function rosenbrock(pop::Population)
     for i=1:length(pop)
         rosenbrock(pop[i])
     end
+    #pop = Population(pmap((x)->rosenbrock(x), pop.chromosomes))
+end
+
+# define the objective function (in-place!)
+function rosenbrock(chr::Chromosome)
+    x1 = chr[1].gene
+    x2 = chr[2].gene
+    chr.fitness = 100*(x2-x1^2)^2+(1-x1)^2-50*((x1+1)^2+(x2-1)^2)-10*((x1-1.5)^2+(x2-2.5)^2)+exp(2*x2-5)
+    chr
 end
 
 # randomly initialize a population with 100 chromosomes

evolib.jl

@@ -269,7 +269,8 @@ rand(T::Type{Chromosome}, num::Int64, x...) = Chromosome([rand(Gene, x...) for i
 
 function rand(T::Type{Chromosome}, num::Int64, obj_func::Function, x...) 
     chr = rand(T, num, x...)
-    chr.fitness = obj_func(chr)
+    #chr.fitness = obj_func(chr)
+    chr = obj_func(chr)
     return chr
 end
 

genetic_test.jl

@@ -9,6 +9,7 @@ function rosenbrock(chr::Chromosome)
     x1 = chr[1].gene
     x2 = chr[2].gene
     chr.fitness = 100*(x2-x1^2)^2+(1-x1)^2-50*((x1+1)^2+(x2-1)^2)-10*((x1-1.5)^2+(x2-2.5)^2)+exp(2*x2-5)
+    return chr
 end
 
 # in case someone wants to calculate the objective function for a population
@@ -26,4 +27,6 @@ probs = GeneticProbabilities(0.2, 0.499, 0.3, 0.001)
 popul = rand(Population, 100, 2, rosenbrock, 0.1, 6.0, -6.0 )
 
 # launch genetic algorithm
-@time genetic(popul, probs, 1000, rosenbrock)
+@time best, conve = genetic(popul, probs, 1000, rosenbrock)
+print(best)
+println()