public static double Fitness(KnapSack a)
{
if (a.Weight > a.MaxWeight)
{
return(0);
}
return(a.Value);
}
public static KnapSack Randomize(double maxWeight, Item[] contents)
{
var newConetents = contents.Select(x => new Item(x.Value, x.Weight, false)).ToArray();
var result = new KnapSack(maxWeight, newConetents);
while (true)
{
var smallEnough = result.Contents.Where(x => result.MaxWeight >= (result.Weight + x.Weight) && !x.InSack).ToArray();
if (!smallEnough.Any())
{
break;
}
var i = rand.Next(0, smallEnough.Length);
smallEnough[i].InSack = true;
}
return(result);
}
static void Main(string[] args)
{
var maxWeight = 20.0;
var baseItems = new[]
{
new KnapSack.Item(10, 5, false),
new KnapSack.Item(5, 2.5, false),
new KnapSack.Item(5, 2.5, false),
new KnapSack.Item(1, 10, false),
new KnapSack.Item(1, 10, false),
new KnapSack.Item(3, 7, false),
new KnapSack.Item(23, 4, false),
new KnapSack.Item(3, 1, false),
new KnapSack.Item(15, 19, false),
new KnapSack.Item(0, 10, false),
new KnapSack.Item(0, 10, false),
new KnapSack.Item(0, 10, false),
new KnapSack.Item(0, 10, false),
new KnapSack.Item(0, 10, false),
new KnapSack.Item(0, 10, false),
new KnapSack.Item(0, 10, false),
new KnapSack.Item(3, 7, false),
new KnapSack.Item(23, 4, false),
new KnapSack.Item(3, 1, false),
new KnapSack.Item(15, 19, false),
new KnapSack.Item(0, 10, false),
new KnapSack.Item(0, 10, false),
new KnapSack.Item(0, 10, false),
new KnapSack.Item(3, 7, false),
new KnapSack.Item(23, 4, false),
new KnapSack.Item(3, 1, false),
new KnapSack.Item(15, 19, false),
new KnapSack.Item(0, 10, false),
new KnapSack.Item(0, 10, false),
new KnapSack.Item(0, 10, false),
};
var pop = Enumerable.Repeat(0, 200).Select(x => KnapSack.Randomize(maxWeight, baseItems)).ToArray();
var fitness = FuncConvert.ToFSharpFunc <KnapSack, double>(KnapSack.Fitness);
var crossover = new Func <KnapSack, KnapSack, KnapSack>(KnapSack.Crossover).ToFSharpFunc();
var mutation = FuncConvert.ToFSharpFunc <KnapSack, KnapSack>(KnapSack.Mutate);
var ga = new GeneticAlgorithm <KnapSack>(pop, fitness, crossover, mutation, .05);
var result = ga.Run();
Console.WriteLine($"Best Fitness: {KnapSack.Fitness(result)} Weight: {result.Weight}");
Console.ReadKey();
}
public static KnapSack Mutate(KnapSack a)
{
var smallEnough = a.Contents.Where(x => !x.InSack && x.Weight + a.Weight <= a.MaxWeight).ToArray();
if (smallEnough.Any())
{
smallEnough[rand.Next(0, smallEnough.Length)].InSack = true;
}
else
{
var inSack = a.Contents.Where(x => x.InSack).ToArray();
if (inSack.Any())
{
inSack[rand.Next(0, inSack.Length)].InSack = false;
}
}
return(a);
}
public static KnapSack Crossover(KnapSack a, KnapSack b)
{
return(new KnapSack(a.MaxWeight, Enumerable.Range(0, a.Contents.Length).Select(i => rand.Next(0, 2) == 0 ? a.Contents[i] : b.Contents[i]).ToArray()));
}