def meta_DE(objective_function,
initial_population,
F,
C,
convergence_tolerance,
convergence_history_length,
max_unconverged_iterations,
mutation_method="rand/1",
selection_method="Storn-Price",
output_function=None):
if 0 < F and F <= 2 and 0 < C and C <= 1:
(iterations, (costs,
final_population)) = differential_evolution.minimize(
objective_function,
initial_population,
F,
C,
convergence_tolerance,
convergence_history_length,
max_unconverged_iterations,
mutation_method=mutation_method,
selection_method=selection_method,
output_function=output_function)
mean_cost = numpy.mean(costs)
print "F = %16.15f, C = %16.15f --> %16.15e after %d iterations." % (
F, C, mean_cost, iterations)
return mean_cost
else:
print "(Constraint violation.)"
return inf
def meta_DE(objective_function, initial_population, F, C,
convergence_tolerance, convergence_history_length, max_unconverged_iterations,
mutation_method="rand/1", selection_method="Storn-Price",
output_function=None):
if 0 < F and F <= 2 and 0 < C and C <= 1:
(iterations, (costs, final_population)) = differential_evolution.minimize(
objective_function, initial_population, F, C,
convergence_tolerance, convergence_history_length, max_unconverged_iterations,
mutation_method=mutation_method, selection_method=selection_method,
output_function=output_function
)
mean_cost = numpy.mean(costs)
print "F = %16.15f, C = %16.15f --> %16.15e after %d iterations." % (F, C, mean_cost, iterations)
return mean_cost
else:
print "(Constraint violation.)"
return inf
return mean_cost
else:
print "(Constraint violation.)"
return inf
def meta_DE_harness(vec):
(F, C) = vec
return meta_DE(
function, numpy.random.uniform(-1.0, 1.0, (50, 9)),
F, C, 0.1, 10, 1000,
mutation_method="MDE5", selection_method="Storn-Price"
)
(iterations, (costs, population)) = differential_evolution.minimize(
meta_DE_harness, numpy.random.uniform(0, 1, (20, 2)),
0.60, 0.90, 0.01, 10, 100,
mutation_method="MDE5", selection_method="elitist",
stochastic=True
)
print "Performed", iterations, "iterations."
best_index = numpy.argmin(costs)
print "Optimized function value:", costs[best_index]
print "Optimized parameter vector:\n", population[best_index]
var88 = abs(var87)
var89 = var86 + var88
var90 = var89**2
return (var4 * var29 + var32 * var37 + var40 * var60 + var63 * var68 +
var71 * var82 + var85 * var90) / 4
t_start = time.clock()
(iterations, (costs, population)) = differential_evolution.minimize(
function,
numpy.random.uniform(-1.0, 1.0, (50, 9)),
0.850,
0.975,
1.0e-12,
250,
12500,
mutation_method="MDE5",
selection_method="Storn-Price",
output_function=None)
t_total = time.clock() - t_start
print "Performed", iterations, "iterations, taking", t_total, "seconds."
numpy.set_printoptions(precision=16, linewidth=109)
best_index = numpy.argmin(costs)
print "Optimized function value:", costs[best_index]
print "Optimized parameter vector:\n", population[best_index]
(F, C) = vec
return meta_DE(function,
numpy.random.uniform(-1.0, 1.0, (50, 9)),
F,
C,
0.1,
10,
1000,
mutation_method="MDE5",
selection_method="Storn-Price")
(iterations, (costs, population)) = differential_evolution.minimize(
meta_DE_harness,
numpy.random.uniform(0, 1, (20, 2)),
0.60,
0.90,
0.01,
10,
100,
mutation_method="MDE5",
selection_method="elitist",
stochastic=True)
print "Performed", iterations, "iterations."
best_index = numpy.argmin(costs)
print "Optimized function value:", costs[best_index]
print "Optimized parameter vector:\n", population[best_index]