def one_run(evaluator, config, frequencies):
'''
Performs a single run of the given configuration. Returns a dictionary
containing results.
Parameters:
- ``evaluator``: An object with the function get_fitness that takes an
individual and returns its fitness value
- ``config``: A dictionary containing all of the configuration information
required to perform a experimental run, including:
- All information required to initialize an individual.
- All information required to run ``evolution.generate``.
- ``verbose``: Boolean value for if extra runtime information should
be displayed.
- ``max_evals``: The maximum number of evaluations allowed before
termination.
- ``max_fitness``: The fitness required to cause a "successful"
termination.
- ``frequencies``: Dictionary used to return information about how often
individuals of different lengths are evolved.
'''
best = Individual(**config)
last_improved = -1
generator = enumerate(generate(config, frequencies))
for evals, individual in generator:
individual.fitness = evaluator.get_fitness(individual)
if best < individual:
best = individual
last_improved = evals
if config['verbose']:
print '\t', last_improved, best.fitness, len(best.active)
if (evals >= config['max_evals']
or best.fitness >= config['max_fitness']):
break
if config['verbose']:
print "Best Found"
best.show_active()
return {
'fitness': best.fitness,
'evals': evals,
'success': best.fitness >= config['max_fitness'],
'phenotype': len(best.active)
}
def one_run(evaluator, config, frequencies):
'''
Performs a single run of the given configuration. Returns a dictionary
containing results.
Parameters:
- ``evaluator``: An object with the function get_fitness that takes an
individual and returns its fitness value
- ``config``: A dictionary containing all of the configuration information
required to perform a experimental run, including:
- All information required to initialize an individual.
- All information required to run ``evolution.generate``.
- ``verbose``: Boolean value for if extra runtime information should
be displayed.
- ``max_evals``: The maximum number of evaluations allowed before
termination.
- ``max_fitness``: The fitness required to cause a "successful"
termination.
- ``frequencies``: Dictionary used to return information about how often
individuals of different lengths are evolved.
'''
best = Individual(**config)
last_improved = -1
generator = enumerate(generate(config, frequencies))
for evals, individual in generator:
individual.fitness = evaluator.get_fitness(individual)
if best < individual:
best = individual
last_improved = evals
if config['verbose']:
print '\t', last_improved, best.fitness, len(best.active)
if (evals >= config['max_evals'] or
best.fitness >= config['max_fitness']):
break
if config['verbose']:
print "Best Found"
best.show_active()
return {'fitness': best.fitness, 'evals': evals,
'success': best.fitness >= config['max_fitness'],
'phenotype': len(best.active)}
def one_run(evaluator, config):
'''
Performs a single run of the given configuration. Returns a dictionary
containing results.
Parameters:
- ``evaluator``: An object with the function get_fitness that takes an
individual and returns its fitness value
- ``config``: A dictionary containing all of the configuration information
required to perform a experimental run, including:
- All information required to initialize an individual.
- All information required to run ``evolution.generate``.
- ``verbose``: Boolean value for if extra runtime information should
be displayed.
- ``max_evals``: The maximum number of evaluations allowed before
termination.
- ``max_fitness``: The fitness required to cause a "successful"
termination.
'''
best = Individual(**config)
output = {}
for evals, individual in enumerate(generate(config, output)):
individual.fitness = evaluator.get_fitness(individual)
if best < individual:
best = individual
if config['verbose']:
print '\t', evals, best.fitness, len(best.active)
if (evals >= config['max_evals'] or
best.fitness >= config['max_fitness']):
break
if config['verbose']:
print "Best Found"
best.show_active()
output.update({'fitness': best.fitness, 'evals': evals,
'success': best.fitness >= config['max_fitness'],
'phenotype': len(best.active),
# When running 'Skip' this will correctly give the number
# of evaluations required by 'Normal'. Otherwise its junk.
'normal': output['skipped'] + evals})
return output