def work(gi, p, s):
G, C, M, k = common.get_stuff(gi)
sample_g = np.linspace(1.45, 1.25, num=p + 2)
sample_g = sample_g[1:p + 1]
sample_b = np.linspace(0.14, 0, num=p + 2)
sample_b = sample_b[1:p + 1]
samples = np.append(sample_g, sample_b)
bounds = [[0, pi / 2] if j < p else [0, pi / 2] for j in range(2 * p)]
kwargs = {'method': 'L-BFGS-B', 'args': (G, C, M, k, p), 'bounds': bounds}
optimal = basinhopping(qaoa,
samples,
minimizer_kwargs=kwargs,
niter=s,
disp=False)
return -optimal.fun
def work(gi, p, s):
G, C, M, k = common.get_stuff(gi)
sample_g, sample_b = common.MLHS(p, s, 0, pi / 2, 0, pi)
bounds = [[0, pi / 2] if j < p else [0, pi] for j in range(2 * p)]
best = -1
for i in range(s):
kwargs = {
'method': 'L-BFGS-B',
'args': (G, C, M, k, p),
'bounds': bounds
}
optimal = basinhopping(qaoa, [sample_g[i], sample_b[i]],
minimizer_kwargs=kwargs,
niter=2,
disp=False)
if -optimal.fun > best: best = -optimal.fun
return best
def work(gi, p, s):
random.seed(random.randint(1,10000) + rank)
G, C, M, k = common.get_stuff(gi)
sample_g, sample_b = common.MLHS(p, s, 0, 0.6, 2.9, pi)
bounds = [[0,0.6] if j < p else [2.9,pi] for j in range(2*p)]
eps = 0.001
best, angles = -1, []
for i in range(s):
kwargs = {'method': 'L-BFGS-B', 'args': (G, C, M, k, p), 'bounds': bounds}
optimal = basinhopping(qaoa, [sample_g[i], sample_b[i]], minimizer_kwargs=kwargs, niter=0, disp=False)
if -optimal.fun > best + eps:
best = -optimal.fun
angles = [list(optimal.x)]
elif -optimal.fun > best - eps:
angles.append(list(optimal.x))
#print(str(rank) + ', ' + str(best) + ', ' + str(angles))
return best, angles
def work(gi, p, s):
G, C, M, k = common.get_stuff(gi)
sample_g, sample_b = common.MLHS(p, s, 0, pi / 2, 0, pi)
bounds = [[0, pi / 2] if j < p else [0, pi] for j in range(2 * p)]
eps = 0.001
best, angles = -1, []
for i in range(s):
kwargs = {
'method': 'L-BFGS-B',
'args': (G, C, M, k, p),
'bounds': bounds
}
optimal = basinhopping(qaoa, [sample_g[i], sample_b[i]],
minimizer_kwargs=kwargs,
niter=2,
disp=False)
if -optimal.fun > best + eps:
best = -optimal.fun
angles = [list(optimal.x)]
elif -optimal.fun > best - eps:
angles.append(list(optimal.x))
return best, angles
def classical(gi, p, s):
G, C, M, k = common.get_stuff(gi)
data = []
num_trials = 20
n_iter = 2
for i in range(s):
print('\ts = ' + str(i) + '\t' + str(datetime.datetime.now().time()))
sample_g, sample_b = common.MLHS(p, num_trials, 0, pi / 2, 0, pi)
init = [[0, pi / 2] if i < p else [0, pi] for i in range(2 * p)]
exp = -1
for i in range(num_trials):
kwargs = {
'method': 'L-BFGS-B',
'args': (G, C, M, k, p),
'bounds': init
}
optimal = basinhopping(qaoa, [sample_g[i], sample_b[i]],
minimizer_kwargs=kwargs,
niter=n_iter,
disp=False)
if -optimal.fun > exp:
exp = -optimal.fun
data.append(exp)
return data