def plot_CIR_edit_types(filename1, filename2, save=False):
# Open the two dataframes
df_1 = open_pickle("data/" + filename1)
df_2 = open_pickle("data/" + filename2)
# Add the edit type to the dataframes
e_type_pf = []
e_type_pp = []
for _ in range(0, df_1.shape[0]):
e_type_pf.append("pertubate_force")
e_type_pp.append("pertubate_points")
df_1.insert(1, "edit_type", e_type_pf, True)
df_2.insert(1, "edit_type", e_type_pp, True)
df = [df_1, df_2]
df = pd.concat(df)
optimal_solution_weight = 9.985281
ax = sns.lineplot(x="n", y="cost", hue="edit_type", data=df)
ax.axhline(optimal_solution_weight,
color="black",
linestyle="dashed",
alpha=0.8)
ax.set(xlabel="n", ylabel="E")
plt.tight_layout()
if save:
plt.savefig("figs/" + filename1 + "_" + filename2 + ".pdf")
else:
plt.show()
def clt():
k = open_pickle("data/clt")
vals = np.split((np.asarray(k) / 10) * 16, 1000)
# Take distribution of one sample
ax = sns.distplot((np.asarray(vals[0]) / 10) * 16, label="mc")
ax = ax.set(xlabel="area", ylabel="frequency")
plt.show()
# Take mean of all samples
ax = sns.distplot(np.mean(vals, axis=1), label="mc")
ax = ax.set(xlabel="area", ylabel="frequency")
plt.show()
def plot_TSP_cs_temp(filename, save=False):
df = open_pickle("data/" + filename + "_cooling_schedules")
ax = sns.lineplot(x="n",
y="temp",
hue="cooling_schedule",
ci=None,
data=df)
plt.tight_layout()
if save:
plt.savefig("figs/" + filename + "_cs_temp_lin.pdf")
else:
plt.show()
plt.clf()
def plot_TSP_temp_temps(filename, save=False):
df = open_pickle("data/" + filename + "_temps")
sns.lineplot(x="n",
y="temp",
hue="init_temp",
ci=None,
data=df,
palette=sns.color_palette('husl', n_colors=3))
plt.tight_layout()
if save:
plt.savefig("figs/" + filename + "_temp_temps.pdf")
else:
plt.show()
plt.clf()
def plot_TSP_mcl(filename, save=False):
df = open_pickle("data/" + filename + "_mcls")
optimal_solution_weight = Network("routes/" +
filename).optimal_solution_weight
ax = sns.lineplot(x="n", y="cost", hue="l", data=df, palette=palette)
ax.axhline(optimal_solution_weight,
color="black",
linestyle="dashed",
alpha=0.8)
plt.tight_layout()
if save:
plt.savefig("figs/" + filename + "_mcls.pdf")
else:
plt.show()
plt.clf()
def plot_CIR_cs_weight(filename, save=False):
df = open_pickle("data/" + filename + "_cooling_schedules")
optimal_solution_weight = 9.985281 # From https://www.nrcresearchpress.com/doi/pdf/10.1139/v88-343
ax = sns.lineplot(x="n", y="cost", hue="cooling_schedule", data=df)
ax.axhline(optimal_solution_weight,
color="black",
linestyle="dashed",
alpha=0.8)
ax.set(xlabel="n", ylabel="E")
plt.tight_layout()
if save:
plt.savefig("figs/" + filename + "_cooling_schedule_cost.pdf")
else:
plt.show()
plt.clf()
def plot_CIR_mcl(filename, save=False):
df = open_pickle("data/" + filename + "_mcls")
optimal_solution_weight = 9.985281
ax = sns.lineplot(x="n", y="cost", hue="l", data=df, palette=palette)
ax.axhline(optimal_solution_weight,
color="black",
linestyle="dashed",
alpha=0.8)
ax.set(xlabel="n", ylabel="E")
plt.tight_layout()
if save:
plt.savefig("figs/" + filename + "mcl" + ".pdf")
plt.ylim(0, 100)
plt.xlim(500, 1000)
plt.savefig("figs/" + filename + "mcl_zoomed" + ".pdf")
else:
plt.show()