This repository has been archived by the owner on Oct 12, 2021. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathprepare_all_benchmark.py
141 lines (120 loc) · 8.27 KB
/
prepare_all_benchmark.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
# Copyright (C) 2020-2021 Adrian Wöltche
#
# This program is free software: you can redistribute it and/or modify
# it under the terms of the GNU Affero General Public License as
# published by the Free Software Foundation, either version 3 of the
# License, or (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU Affero General Public License for more details.
#
# You should have received a copy of the GNU Affero General Public License
# along with this program. If not, see https://www.gnu.org/licenses/.
import os
import pandas as pd
def export_column(dataframe, column, filepattern, sort=False, reverse=False, percentile=False):
if type(dataframe) == pd.Series:
tmp = dataframe
else:
tmp = dataframe[column]
if sort:
tmp = tmp.sort_values(ascending=not reverse).reset_index()
tmp['index'] = tmp.index
if percentile:
tmp['index'] = tmp['index'] / tmp['index'].max()
tmp = tmp.rename(columns={'index': 'x', column: 'y'})
tmp.to_csv(filepattern.format(column), index=False)
export_pattern = os.path.join('benchmarks', 'benchmark_{}.csv')
df = pd.read_csv("benchmark_all.csv")
count = df.count()[0]
def error_row(name, column):
print(name,
"&", "${:.2f} \\%$".format((df[df[column] <= 0.0].count()[0] / count * 100)),
"&", "${:.2f}$".format(df[df[column] < float('inf')][column].mean()),
"&", "${:.2f}$".format(df[df[column] < float('inf')][column].quantile(0.90)),
"&", "${:.2f}$".format(df[df[column] < float('inf')][column].quantile(0.95)),
"&", "${:.2f}$".format(df[df[column] < float('inf')][column].quantile(0.99)),
"\\\\")
def duration_row(name, base, column):
tmp = df[base] / df[column]
print(name,
"&", "${:.2f} \\%$".format((df[tmp > 1.0].count()[0] / count * 100)),
"&", "${:.2f}$".format(tmp.quantile(0.01)),
"&", "${:.2f}$".format(tmp.quantile(0.25)),
"&", "${:.2f}$".format(tmp.median()),
"&", "${:.2f}$".format(tmp.quantile(0.75)),
"&", "${:.2f}$".format(tmp.quantile(0.99)),
"&", "${:.2f}$".format(tmp.mean()),
"\\\\")
def duration_state_space(name, column):
print(name,
"&", "${:.2f}$".format(df[column].quantile(0.01)),
"&", "${:.2f}$".format(df[column].quantile(0.10)),
"&", "${:.2f}$".format(df[column].quantile(0.25)),
"&", "${:.2f}$".format(df[column].median()),
"&", "${:.2f}$".format(df[column].quantile(0.75)),
"&", "${:.2f}$".format(df[column].quantile(0.90)),
"&", "${:.2f}$".format(df[column].quantile(0.99)),
"&", "${:.2f}$".format(df[column].mean()),
"\\\\")
print("\nError fraction table\n")
error_row("Value Iteration", 'value_iteration_error_fraction')
error_row("Q-Learning $\\epsilon$-greedy", 'qlearning_epsilon_error_fraction')
error_row("Q-Learning $\\epsilon$-decay", 'qlearning_epsilon_decay_error_fraction')
error_row("Q-Learning* Intelligent", 'qlearning_intelligent_error_fraction')
error_row("Q-Learning* Greedy", 'qlearning_greedy_error_fraction')
error_row("Expected SARSA", 'expected_sarsa_error_fraction')
error_row("Greedy", 'greedy_error_fraction')
error_row("Nearest", 'nearest_error_fraction')
print("\nDuration table\n")
duration_row("Value Iteration", 'value_iteration_duration', 'value_iteration_duration')
duration_row("Q $\\epsilon$-greedy", 'value_iteration_duration', 'qlearning_epsilon_duration')
duration_row("Q $\\epsilon$-decay", 'value_iteration_duration', 'qlearning_epsilon_decay_duration')
duration_row("Q* Intelligent", 'value_iteration_duration', 'qlearning_intelligent_duration')
duration_row("Q* Greedy", 'value_iteration_duration', 'qlearning_greedy_duration')
duration_row("Expected SARSA", 'value_iteration_duration', 'expected_sarsa_duration')
duration_row("Greedy", 'value_iteration_duration', 'greedy_duration')
duration_row("Nearest", 'value_iteration_duration', 'nearest_duration')
print("\nState space exploration table\n")
duration_state_space("Value Iteration", 'value_iteration_states_percentage')
duration_state_space("Q-Learning $\\epsilon$-greedy", 'qlearning_epsilon_states_percentage')
duration_state_space("Q-Learning $\\epsilon$-decay", 'qlearning_epsilon_decay_states_percentage')
duration_state_space("Q-Learning* Intelligent", 'qlearning_intelligent_states_percentage')
duration_state_space("Q-Learning* Greedy", 'qlearning_greedy_states_percentage')
duration_state_space("Expected SARSA", 'expected_sarsa_states_percentage')
duration_state_space("Greedy", 'greedy_states_percentage')
duration_state_space("Nearest", 'nearest_states_percentage')
# error fraction
export_column(df, 'value_iteration_error_fraction', export_pattern, sort=True, reverse=True, percentile=True)
export_column(df, 'nearest_error_fraction', export_pattern, sort=True, reverse=True, percentile=True)
export_column(df, 'greedy_error_fraction', export_pattern, sort=True, reverse=True, percentile=True)
export_column(df, 'qlearning_intelligent_error_fraction', export_pattern, sort=True, reverse=True, percentile=True)
export_column(df, 'qlearning_greedy_error_fraction', export_pattern, sort=True, reverse=True, percentile=True)
export_column(df, 'qlearning_epsilon_error_fraction', export_pattern, sort=True, reverse=True, percentile=True)
export_column(df, 'qlearning_epsilon_decay_error_fraction', export_pattern, sort=True, reverse=True, percentile=True)
export_column(df, 'expected_sarsa_error_fraction', export_pattern, sort=True, reverse=True, percentile=True)
export_column(df, 'expected_sarsa_decay_error_fraction', export_pattern, sort=True, reverse=True, percentile=True)
# duration
export_column(df['value_iteration_duration'] / df['value_iteration_duration'], 'value_iteration_duration', export_pattern, sort=True, reverse=False, percentile=True)
export_column(df['value_iteration_duration'] / df['nearest_duration'], 'nearest_duration', export_pattern, sort=True, reverse=False, percentile=True)
export_column(df['value_iteration_duration'] / df['greedy_duration'], 'greedy_duration', export_pattern, sort=True, reverse=False, percentile=True)
export_column(df['value_iteration_duration'] / df['qlearning_intelligent_duration'], 'qlearning_intelligent_duration', export_pattern, sort=True, reverse=False, percentile=True)
export_column(df['value_iteration_duration'] / df['qlearning_greedy_duration'], 'qlearning_greedy_duration', export_pattern, sort=True, reverse=False, percentile=True)
export_column(df['value_iteration_duration'] / df['qlearning_epsilon_duration'], 'qlearning_epsilon_duration', export_pattern, sort=True, reverse=False, percentile=True)
export_column(df['value_iteration_duration'] / df['qlearning_epsilon_decay_duration'], 'qlearning_epsilon_decay_duration', export_pattern, sort=True, reverse=False, percentile=True)
export_column(df['value_iteration_duration'] / df['expected_sarsa_duration'], 'expected_sarsa_duration', export_pattern, sort=True, reverse=False, percentile=True)
export_column(df['value_iteration_duration'] / df['expected_sarsa_decay_duration'], 'expected_sarsa_decay_duration', export_pattern, sort=True, reverse=False, percentile=True)
# state space
export_column(df, 'value_iteration_states_percentage', export_pattern, sort=True, reverse=True, percentile=True)
export_column(df, 'nearest_states_percentage', export_pattern, sort=True, reverse=True, percentile=True)
export_column(df, 'greedy_states_percentage', export_pattern, sort=True, reverse=True, percentile=True)
export_column(df, 'qlearning_intelligent_states_percentage', export_pattern, sort=True, reverse=True, percentile=True)
export_column(df, 'qlearning_greedy_states_percentage', export_pattern, sort=True, reverse=True, percentile=True)
export_column(df, 'qlearning_epsilon_states_percentage', export_pattern, sort=True, reverse=True, percentile=True)
export_column(df, 'qlearning_epsilon_decay_states_percentage', export_pattern, sort=True, reverse=True, percentile=True)
export_column(df, 'expected_sarsa_states_percentage', export_pattern, sort=True, reverse=True, percentile=True)
export_column(df, 'expected_sarsa_decay_states_percentage', export_pattern, sort=True, reverse=True, percentile=True)