// Solve returns a Magic representing the solution to a magic square schema. The
// schema is a partially filled magic square, interpreted as the concatination
// of the rows of the square. Values greater than or equal to 0 appearing in the
// schema must appear in the solution at the same position, and values less than
// 0 in the schema represent unknown values.
//
// The value is returned on a channel allocated by the caller.
//
// The population is reseeded occasionally if the solution is not found. The
// return value rsc is the "reseed count" giving the number of times reseeding
// occurs.
func Solve(schema Square, ret chan Square) {
var (
// the number of values in the square
order = len(schema.Sq)
// the size of the population
popz = 10 * order
// the interface into the GA loop
pop evo.Population
// the initial population
seed = make([]evo.Genome, popz)
)
// Print the population when receiving a SIGQUIT.
// (press ctrl-/ at the terminal)
go func() {
sigs := make(chan os.Signal, 1)
signal.Notify(sigs, syscall.SIGQUIT)
<-sigs
stats := pop.Stats()
fmt.Println(stats.SD(), stats.RSD())
for i := pop.Iter(); i.Value() != nil; i.Next() {
fmt.Println(i.Value().(*Genome).Express(), i.Value().(*Genome).Fitness())
}
os.Exit(2)
}()
// start the GA in the background
for i := range seed {
seed[i] = &Genome{RandSiam(schema)}
}
pop = gen.New(seed)
pop.Start()
for {
best := evo.Max(pop)
stats := pop.Stats()
// halt when we find a solution
if best.Fitness() == 0 {
ret <- best.(*Genome).Square
pop.Close()
return
}
// reseed
if stats.RSD() < 0.3 {
pop.Close()
for i := range seed {
seed[i] = &Genome{RandSiam(schema)}
}
seed[0] = best
pop = gen.New(seed)
pop.Start()
}
}
}
func TestAckley(t *testing.T) {
fmt.Printf("Minimize the Ackley function with n=%d\n", dim)
// Setup:
// We initialize a set of 40 random solutions,
// then add them to a generational population.
init := make([]evo.Genome, 40)
for i := range init {
init[i] = &ackley{
gene: real.Random(dim, 30),
steps: real.Random(dim, 1),
}
}
pop := gen.New(init)
pop.Start()
// Tear-down:
// Upon returning, we cleanup our resources and print the solution.
defer func() {
pop.Close()
selector.Close()
fmt.Println("\nSolution:", evo.Max(pop))
}()
// Run:
// We continuously poll the population for statistics and terminate when we
// have a solution or after 200,000 evaluations.
for {
count.Lock()
n := count.n
count.Unlock()
stats := pop.Stats()
// "\x1b[2K" is the escape code to clear the line
// The fitness of minimization problems is negative
fmt.Printf("\x1b[2K\rCount: %7d | Max: %8.3g | Mean: %8.3g | Min: %8.3g | RSD: %9.2e",
n,
-stats.Min(),
-stats.Mean(),
-stats.Max(),
stats.RSD())
// We've converged once the deviation is within the precision
if stats.SD() < precision {
return
}
// Force stop after 200,000 fitness evaluations
if n > 200000 {
return
}
}
}
func TestTSP(t *testing.T) {
fmt.Println("Minimize tour of US capitals - optimal is", best)
// Setup:
// We create a random initial population
// and evolve it using a generational model.
init := make([]evo.Genome, size)
for i := range init {
init[i] = &tsp{gene: pool.Get().([]int)}
}
pop = graph.Hypercube(init)
pop.Start()
// Tear-down:
// Upon returning, we cleanup our resources and print the solution.
defer func() {
pop.Close()
fmt.Println("\nTour:", evo.Max(pop))
}()
// Run:
// We continuously poll the population for statistics used in the
// termination conditions.
for {
count.Lock()
n := count.n
count.Unlock()
stats := pop.Stats()
// "\x1b[2K" is the escape code to clear the line
// The fitness of minimization problems is negative
fmt.Printf("\x1b[2K\rCount: %7d | Max: %6.0f | Mean: %6.0f | Min: %6.0f | RSD: %7.2e",
n,
-stats.Min(),
-stats.Mean(),
-stats.Max(),
stats.RSD())
// Stop when we get close. Finding the true minimum could take a while.
if -stats.Max() < best*1.1 {
return
}
// Force stop after some number fitness evaluations
if n > stop {
t.Fail()
return
}
}
}
func TestQueens(t *testing.T) {
fmt.Printf("Find a solution to %d-queens\n", dim)
// Setup:
// We create a random initial population and divide it into islands. Each
// island is evolved independently. The islands are grouped together into
// a wrapping population. The wrapper coordiates migrations between the
// islands according to the delay period.
init := make([]evo.Genome, size)
for i := range init {
init[i] = &queens{gene: pool.Get().([]int)}
}
islands := make([]evo.Genome, isl)
islSize := size / isl
for i := range islands {
islands[i] = gen.New(init[i*islSize : (i+1)*islSize])
islands[i].(evo.Population).Start()
}
pop := graph.Ring(islands)
pop.SetDelay(delay)
pop.Start()
// Tear-down:
// Upon returning, we cleanup our resources and print the solution.
defer func() {
pop.Close()
fmt.Println("\nSolution:", evo.Max(pop))
}()
// Run:
// We continuously poll the population for statistics used in the
// termination conditions.
for {
count.Lock()
n := count.n
count.Unlock()
stats := pop.Stats()
// "\x1b[2K" is the escape code to clear the line
// The fitness of minimization problems is negative
fmt.Printf("\x1b[2K\rCount: %7d | Max: %3.0f | Mean: %3.0f | Min: %3.0f | RSD: %9.2e",
n,
-stats.Min(),
-stats.Mean(),
-stats.Max(),
stats.RSD())
// We've found the solution when max is 0
if stats.Max() == 0 {
return
}
// We've converged once the deviation is less than 0.01
if stats.SD() < 1e-2 {
return
}
// Force stop after 2,000,000 fitness evaluations
if n > 2e6 {
return
}
// var blocker chan struct{}
// <-blocker
}
}