def _calculate(benchmark_name, dimensionality, verbose):
"""Calculate the requested measurement."""
benchmark = benchmarks.get(benchmark_name)
f = benchmark.function
f_min = benchmark.min(0)
f_max = benchmark.max(0)
return ruggedness.FEM_0_1(f, f_min, f_max, dimensionality, verbose=verbose)
def _calculate(benchmark_name, dimensionality, verbose=False):
"""Calculate the requested measurement."""
benchmark = benchmarks.get(benchmark_name)
f = benchmark.function
f_min = benchmark.min(0)
f_max = benchmark.max(0)
return funnels.DM(f, f_min, f_max, dimensionality, verbose=verbose)
def _calculate(benchmark_name, dimensionality):
"""Calculate the requested measurement."""
benchmark = benchmarks.get(benchmark_name)
f = benchmark.function
f_min = benchmark.min(0)
f_max = benchmark.max(0)
return searchability.FCI_soc(f, f_min, f_max, dimensionality)
def _calculate(benchmark_name, dimensionality):
"""Calculate the requested measurement."""
benchmark = benchmarks.get(benchmark_name)
f = benchmark.function
f_min = benchmark.min(0)
f_max = benchmark.max(0)
return deception.FDC(f, f_min, f_max, dimensionality)
def _calculate(benchmark_name, dimensionality, step_size_fraction):
"""Calculate the requested measurement."""
benchmark = benchmarks.get(benchmark_name)
f = benchmark.function
f_min = benchmark.min(0)
f_max = benchmark.max(0)
g_avg, g_dev = gradients.G_measures(f,
f_min,
f_max,
dimensionality,
step_size_fraction=step_size_fraction)
return g_avg, g_dev
def process(batch_num, num_batches, verbose):
config = _config(batch_num)
while config is not None:
benchmark_name, dimensionality, experiment = config
benchmark = benchmarks.get(benchmark_name)
if benchmark.is_dimensionality_valid(dimensionality):
print 'DM: getting', benchmark_name, dimensionality, experiment
dm.get(benchmark_name, dimensionality, experiment, verbose=verbose)
else:
print 'DM: skipping', benchmark_name, dimensionality, experiment, '(invalid number of dimensions)'
batch_num += num_batches
config = _config(batch_num)
def _calculate(benchmark_name, dimensionality, epsilon, step_size_fraction):
"""Calculate the requested measurement."""
benchmark = benchmarks.get(benchmark_name)
f = benchmark.function
f_min = benchmark.min(0)
f_max = benchmark.max(0)
pn, lsn = neutrality.PN_LSN(f,
f_min,
f_max,
dimensionality,
epsilon=epsilon,
step_size_fraction=step_size_fraction)
return pn, lsn
def process(batch_num, num_batches, verbose):
config = _config(batch_num)
while config is not None:
benchmark_name, dimensionality, experiment, is_soc = config
benchmark = benchmarks.get(benchmark_name)
if benchmark.is_dimensionality_valid(dimensionality):
if is_soc:
print 'FCI_soc: getting', benchmark_name, dimensionality, experiment
# fci_sigma.get(benchmark_name, dimensionality, verbose=verbose)
fci_soc.get(benchmark_name, dimensionality,
experiment, verbose=verbose)
else:
print 'FCI_cog: getting', benchmark_name, dimensionality, experiment
fci_cog.get(benchmark_name, dimensionality,
experiment, verbose=verbose)
else:
print 'FCI: skipping', benchmark_name, dimensionality, '(invalid number of dimensions)'
batch_num += num_batches
config = _config(batch_num)
