def pp_lmfit_fitting(plan, real_perf_values, method='leastsq'):
'''least squares or differential evolution fitting with bounds'''
real_perf_values = np.array(real_perf_values)
load_scale_factor = calc_pp_load_scale_factor(plan)
perf_scale_factor = calc_pp_perf_scale_factor(real_perf_values)
scaled_plan = list(map(lambda l: load_scale_factor * l, plan))
scaled_perfs = list(map(lambda p: perf_scale_factor * p, real_perf_values))
args = (scaled_plan,
np.array(scaled_perfs),
calc_residuals,
unpack_pp_lmfit_parms,
pp_performance_over_time2)
params = lmfit.Parameters()
params.add(name='perfpot', value=0.5, min=0, max=1)
params.add(name='straindelay', value=4.0, min=0.001, max=30)
params.add(name='responsedelay', value=2.0, min=0.001, max=30)
params.add(name='overflowdelay', value=15, min=0.001, max=30)
lmfit.minimize(objective_f, params, method=method, args=args)
model_perfs = pp_performance_over_time(scaled_plan,
0.0,
0.0,
params['perfpot'],
params['straindelay'],
params['responsedelay'],
params['overflowdelay'])
model_perfs = filter_model_perfs_2_real_perfs(model_perfs, real_perf_values)
scaled_perfs = list(filter(lambda x: x > 0.0, scaled_perfs))
assert(len(model_perfs) == len(scaled_perfs))
rmse = calc_rmse(scaled_perfs, model_perfs)
return (((params['perfpot'].value,
params['straindelay'].value,
params['responsedelay'].value,
params['overflowdelay'].value), rmse),
load_scale_factor,
perf_scale_factor)
def pp_cmaes_fitting(plan, real_perf_values):
# x0 = np.array([4.0, 2.0, 15]) # initial guess of delays
x0 = np.array([0.5, 4.0, 2.0, 15]) # initial guess including perfpot
load_scale_factor = calc_pp_load_scale_factor(plan)
perf_scale_factor = calc_pp_perf_scale_factor(real_perf_values)
scaled_plan = list(map(lambda l: load_scale_factor * l, plan))
scaled_perfs = list(map(lambda p: perf_scale_factor * p, real_perf_values))
args = (scaled_plan,
scaled_perfs,
calc_rmse,
unpack_pp_parms_list,
pp_performance_over_time2)
opts = cma.CMAOptions()
# only optimize delays
'''bounds = [[0.001, 0.001, 0.001],
[30.0, 30.0, 30.0]]'''
# optimize delays and init_p aka perfpot
bounds = [[0.0, 0.001, 0.001, 0.001],
[1.0, 30.0, 30.0, 30.0]]
opts.set('bounds', bounds)
# opts.set('verb_disp', False)
# opts.set('verbose', -9)
opts.set('maxiter', 800)
res = cma.fmin(objective_f, x0, 0.5, args=args, options=opts)
print('res[0] = {}'.format(res[0]))
print('res[1] = {}'.format(res[1]))
return res, load_scale_factor, perf_scale_factor
def pp_minimize_fitting(plan, real_perf_values, method):
'''generic interface for optimization.minimize for PerPot fitting.
returns the OptimizeResult object and the scale factors.'''
# x0 = np.array([4.0, 2.0, 0.001]) # initial guess
x0 = np.array([0.5, 4.0, 2.0, 15]) # initial guess including perfpot
load_scale_factor = calc_pp_load_scale_factor(plan)
perf_scale_factor = calc_pp_perf_scale_factor(real_perf_values)
scaled_plan = list(map(lambda l: load_scale_factor * l, plan))
scaled_perfs = list(map(lambda p: perf_scale_factor * p, real_perf_values))
args = (scaled_plan,
scaled_perfs,
calc_rmse,
unpack_pp_parms_list,
pp_performance_over_time2)
bounds = [(0.0, 1.0), # perfpot
(0.001, 30.0), # DS Delay of Strain Rate
(0.001, 30.0), # DR Delay of Response Rate
(0.001, 30.0)] # DSO Delay of Strain Overflow Rate
options = {}
# options['maxiter'] = 50
options['disp'] = False
def iter_callback(xk):
print("current parameter vector: {}".format(xk))
optres = optimize.minimize(objective_f,
x0,
args,
method,
bounds=bounds,
options=options,
callback=iter_callback)
return optres, load_scale_factor, perf_scale_factor
def pp_lmfit_fitting(plan, real_perf_values, method='leastsq'):
'''least squares or differential evolution fitting with bounds'''
real_perf_values = np.array(real_perf_values)
load_scale_factor = calc_pp_load_scale_factor(plan)
perf_scale_factor = calc_pp_perf_scale_factor(real_perf_values)
scaled_plan = list(map(lambda l: load_scale_factor * l, plan))
scaled_perfs = list(map(lambda p: perf_scale_factor * p, real_perf_values))
args = (scaled_plan, np.array(scaled_perfs), calc_residuals,
unpack_pp_lmfit_parms, pp_performance_over_time2)
params = lmfit.Parameters()
params.add(name='perfpot', value=0.5, min=0, max=1)
params.add(name='straindelay', value=4.0, min=0.001, max=30)
params.add(name='responsedelay', value=2.0, min=0.001, max=30)
params.add(name='overflowdelay', value=15, min=0.001, max=30)
lmfit.minimize(objective_f, params, method=method, args=args)
model_perfs = pp_performance_over_time(scaled_plan, 0.0, 0.0,
params['perfpot'],
params['straindelay'],
params['responsedelay'],
params['overflowdelay'])
model_perfs = filter_model_perfs_2_real_perfs(model_perfs,
real_perf_values)
scaled_perfs = list(filter(lambda x: x > 0.0, scaled_perfs))
assert (len(model_perfs) == len(scaled_perfs))
rmse = calc_rmse(scaled_perfs, model_perfs)
return (((params['perfpot'].value, params['straindelay'].value,
params['responsedelay'].value, params['overflowdelay'].value),
rmse), load_scale_factor, perf_scale_factor)
def pp_minimize_fitting(plan, real_perf_values, method):
'''generic interface for optimization.minimize for PerPot fitting.
returns the OptimizeResult object and the scale factors.'''
# x0 = np.array([4.0, 2.0, 0.001]) # initial guess
x0 = np.array([0.5, 4.0, 2.0, 15]) # initial guess including perfpot
load_scale_factor = calc_pp_load_scale_factor(plan)
perf_scale_factor = calc_pp_perf_scale_factor(real_perf_values)
scaled_plan = list(map(lambda l: load_scale_factor * l, plan))
scaled_perfs = list(map(lambda p: perf_scale_factor * p, real_perf_values))
args = (scaled_plan, scaled_perfs, calc_rmse, unpack_pp_parms_list,
pp_performance_over_time2)
bounds = [
(0.0, 1.0), # perfpot
(0.001, 30.0), # DS Delay of Strain Rate
(0.001, 30.0), # DR Delay of Response Rate
(0.001, 30.0)
] # DSO Delay of Strain Overflow Rate
options = {}
# options['maxiter'] = 50
options['disp'] = False
def iter_callback(xk):
print("current parameter vector: {}".format(xk))
optres = optimize.minimize(objective_f,
x0,
args,
method,
bounds=bounds,
options=options,
callback=iter_callback)
return optres, load_scale_factor, perf_scale_factor
def pp_cmaes_fitting(plan, real_perf_values):
# x0 = np.array([4.0, 2.0, 15]) # initial guess of delays
x0 = np.array([0.5, 4.0, 2.0, 15]) # initial guess including perfpot
load_scale_factor = calc_pp_load_scale_factor(plan)
perf_scale_factor = calc_pp_perf_scale_factor(real_perf_values)
scaled_plan = list(map(lambda l: load_scale_factor * l, plan))
scaled_perfs = list(map(lambda p: perf_scale_factor * p, real_perf_values))
args = (scaled_plan, scaled_perfs, calc_rmse, unpack_pp_parms_list,
pp_performance_over_time2)
opts = cma.CMAOptions()
# only optimize delays
'''bounds = [[0.001, 0.001, 0.001],
[30.0, 30.0, 30.0]]'''
# optimize delays and init_p aka perfpot
bounds = [[0.0, 0.001, 0.001, 0.001], [1.0, 30.0, 30.0, 30.0]]
opts.set('bounds', bounds)
# opts.set('verb_disp', False)
# opts.set('verbose', -9)
opts.set('maxiter', 800)
res = cma.fmin(objective_f, x0, 0.5, args=args, options=opts)
print('res[0] = {}'.format(res[0]))
print('res[1] = {}'.format(res[1]))
return res, load_scale_factor, perf_scale_factor
