def run_steepest_func_2():
csv_table = CSVFormatter()
# x_0 * x_1 + (1-x_0)^2 + (2-x_1)^2+ (3-x_2)^2+ (4-x_3)^2+ (5-x_4)^2+ (6-x_5)^2+ (7-x_6)^2+ (8-x_7)^2+ (9-x_8)^2+ (10-x_9)^2
## 10 Dimension
test_func = lambda x: x[0] * x[1] + sum([(ele - x[ele - 1])**2
for ele in range(1, 11)])
fprime = lambda x: [x[1] + -2 * (1 - x[0]),
x[0] + -2 * (2 - x[1])] + \
[-2*(ele - x[ele - 1]) for ele in range(3, 11)]
csv_table.add_row(STEEP_HEADER)
x0_param = [[random.gauss(-10, 10) for _ in range(10)]
for _ in range(NUM_RUNS)]
eps = 1e-14
for i in range(10):
gc_iterations_param_dist = [10] * NUM_RUNS
epsilon_abs_param_dist = [eps] * NUM_RUNS
run_report_list = TestHarness.test_optimizer(
alternate_method, test_func, x0_param,
[fprime for _ in range(NUM_RUNS)], epsilon_abs_param_dist)
stats = StatsGenerator.generate_stats(run_report_list, (1, 1), 1)
print(stats)
csv_table.add_row([
epsilon_abs_param_dist[i],
CELL_FORMAT_STR.format(stats[0][0], stats[0][1]),
CELL_FORMAT_STR.format(stats[1][0], stats[1][1])
])
eps *= 10
csv_table.export_table("steepest_func2.csv")
print("Finished steepest decent function 2")
csv_table.clear_table()
def run_steepest_func_1():
csv_table = CSVFormatter()
## FIRST FUNCTION
## Rosenblock
test_func = lambda x: (1 - x[0])**2 + 100 * (x[1] - x[0]**2)**2 + 1
fprime = lambda x: (2 *
(200 * x[0]**3 - 200 * x[0] * x[1] + x[0] - 1), 200 *
(x[1] - x[0]**2))
csv_table.add_row(STEEP_HEADER)
# Func 1 epsilon variation
x0_param = [(random.gauss(-10, 10), random.gauss(-10, 10))
for _ in range(NUM_RUNS)]
eps = 1e-14
for i in range(10):
gc_iterations_param_dist = [10] * NUM_RUNS
epsilon_abs_param_dist = [eps] * NUM_RUNS
run_report_list = TestHarness.test_optimizer(
alternate_method, test_func, x0_param,
[fprime for _ in range(NUM_RUNS)], epsilon_abs_param_dist)
stats = StatsGenerator.generate_stats(run_report_list, (1, 1), 1)
print(stats)
csv_table.add_row([
epsilon_abs_param_dist[i],
CELL_FORMAT_STR.format(stats[0][0], stats[0][1]),
CELL_FORMAT_STR.format(stats[1][0], stats[1][1])
])
eps *= 10
csv_table.export_table("steepest_func1.csv")
print("Finished steepest decent function 1")
csv_table.clear_table()
def problem_1_b():
csvf = CSVFormatter()
csvf.add_row(["Time (ms)"])
color_set = range(3)
problem_size = 10
num_runs = 100
cg = GraphColoring(color_set)
harness_param_x0 = [[
random.choice(color_set) for _ in range(problem_size)
] for _ in range(num_runs)]
harness_allowed = [[color_set for _ in range(problem_size)]] * num_runs
harness_t0 = [25000] * num_runs
harness_t_final = [0.1] * num_runs
run_report_list = TestHarness.test_optimizer(simulated_annealing, cg.solve,
harness_param_x0,
harness_allowed, harness_t0,
harness_t_final)
print("Finished coloring graph runs")
time_stats = StatsGenerator.stats_to_string(
StatsGenerator.generate_stats(
list(map(lambda x: x[0], run_report_list))))
completion_stats = StatsGenerator.stats_to_string(
StatsGenerator.generate_stats(
list(map(lambda x: x[2], run_report_list))))
print(time_stats, completion_stats)
csvf.add_row([time_stats, completion_stats])
csvf.export_table("sa_graph_coloring")
def problem_2():
nm_csvf = CSVFormatter()
sa_csvf = CSVFormatter()
nm_csvf.add_row(["Time (ms)", "Minimum"])
sa_csvf.add_row(["Time (ms)", "Minimum"])
nm_test_function = lambda x: x[0]**2 + 2 * x[1]**2 + 2 * x[0] * x[1]
num_runs = 1000
harness_sa_allowed = [[[], []]] * num_runs
harness_sa_t0 = [25000] * num_runs
harness_sa_t_final = [0.1] * num_runs
harness_x0 = [(random.uniform(-100, 100), random.uniform(-100, 100))
for _ in range(num_runs)]
nm_run_report = TestHarness.test_optimizer(nelder_mead, nm_test_function,
harness_x0)
sa_run_report = TestHarness.test_optimizer(simulated_annealing,
nm_test_function, harness_x0,
harness_sa_allowed,
harness_sa_t0,
harness_sa_t_final)
nm_time_stats = StatsGenerator.generate_stats(
list(map(lambda x: x[0], nm_run_report)))
nm_min_stats = StatsGenerator.generate_stats(
list(map(lambda x: x[2], nm_run_report)))
sa_time_stats = StatsGenerator.generate_stats(
list(map(lambda x: x[0], sa_run_report)))
sa_min_stats = StatsGenerator.generate_stats(
list(map(lambda x: x[2], sa_run_report)))
nm_csvf.add_row([
StatsGenerator.stats_to_string(nm_time_stats),
StatsGenerator.stats_to_string(nm_min_stats)
])
sa_csvf.add_row([
StatsGenerator.stats_to_string(sa_time_stats),
StatsGenerator.stats_to_string(sa_min_stats)
])
nm_csvf.export_table("tables/nelder_mead_stats.csv")
sa_csvf.export_table("tables/continuous_simulated_annealing_stats.csv")
def run_cg_func_2():
csv_table = CSVFormatter()
test_func = lambda x: x[0] * x[1] + sum([(ele - x[ele - 1])**2
for ele in range(1, 11)])
fprime = lambda x: [x[1] + -2 * (1 - x[0]),
x[0] + -2 * (2 - x[1])] + \
[-2*(ele - x[ele - 1]) for ele in range(3, 11)]
x0_param = [[random.gauss(-10, 10) for _ in range(10)]
for _ in range(NUM_RUNS)]
epsilon_abs_param_dist = [1.14e-14] * NUM_RUNS
csv_table.add_row(CG_HEADER)
for i in range(NUM_VARIATIONS):
gc_iterations_param_dist = [random.randint(2, NUM_VARIATIONS)
] * NUM_RUNS
run_report_list = TestHarness.test_optimizer(
conjugate_gradients, test_func, x0_param,
[fprime for _ in range(NUM_RUNS)], gc_iterations_param_dist,
epsilon_abs_param_dist)
stats = StatsGenerator.generate_stats(run_report_list,
(0, 2, 3, 4, 5, 6, 7, 8, 9, 10),
1)
print(stats)
csv_table.add_row([
gc_iterations_param_dist[i], epsilon_abs_param_dist[i],
CELL_FORMAT_STR.format(stats[0][0], stats[0][1]),
CELL_FORMAT_STR.format(stats[1][0], stats[1][1]),
CELL_FORMAT_STR.format(stats[2][0], stats[2][1])
])
csv_table.export_table("function2_gc_freq_variation.csv")
csv_table.clear_table()
print("Finished function 2 GC restart variation")
csv_table.add_row(CG_HEADER)
# Func 1 epsilon variation
eps = 1e-14
for i in range(10):
gc_iterations_param_dist = [10] * NUM_RUNS
epsilon_abs_param_dist = [eps] * NUM_RUNS
run_report_list = TestHarness.test_optimizer(
conjugate_gradients, test_func, x0_param,
[fprime for _ in range(NUM_RUNS)], gc_iterations_param_dist,
epsilon_abs_param_dist)
stats = StatsGenerator.generate_stats(run_report_list,
(0, 2, 3, 4, 5, 6, 7, 8, 9, 10),
1)
print(stats)
csv_table.add_row([
gc_iterations_param_dist[i], epsilon_abs_param_dist[i],
CELL_FORMAT_STR.format(stats[0][0], stats[0][1]),
CELL_FORMAT_STR.format(stats[1][0], stats[1][1]),
CELL_FORMAT_STR.format(stats[2][0], stats[2][1])
])
eps *= 10
csv_table.export_table("function2_epsilon_variation.csv")
print("Finished function 2 epsilon variation")
csv_table.clear_table()
def run_cg_func_1():
csv_table = CSVFormatter()
## FIRST FUNCTION
## Rosenblock
test_func = lambda x: (1 - x[0])**2 + 100 * (x[1] - x[0]**2)**2 + 1
fprime = lambda x: (2 *
(200 * x[0]**3 - 200 * x[0] * x[1] + x[0] - 1), 200 *
(x[1] - x[0]**2))
# Vary GC restart frequency (2->50 drawn from uniform distribution
# Func 1 conj variation
epsilon_abs_param_dist = [1.14e-14] * NUM_RUNS
x0_param = [(random.gauss(-10, 10), random.gauss(-10, 10))
for _ in range(NUM_RUNS)]
csv_table.add_row(CG_HEADER)
for i in range(NUM_VARIATIONS):
gc_iterations_param_dist = [random.randint(2, NUM_VARIATIONS)
] * NUM_RUNS
run_report_list = TestHarness.test_optimizer(
conjugate_gradients, test_func, x0_param,
[fprime for _ in range(NUM_RUNS)], gc_iterations_param_dist,
epsilon_abs_param_dist)
stats = StatsGenerator.generate_stats(run_report_list, (1, 1), 1)
print(stats)
csv_table.add_row([
gc_iterations_param_dist[i], epsilon_abs_param_dist[i],
CELL_FORMAT_STR.format(stats[0][0], stats[0][1]),
CELL_FORMAT_STR.format(stats[1][0], stats[1][1]),
CELL_FORMAT_STR.format(stats[2][0], stats[2][1])
])
csv_table.export_table("function1_gc_freq_variation.csv")
csv_table.clear_table()
print("Finished function 1 GC restart variation")
csv_table.add_row(CG_HEADER)
# Func 1 epsilon variation
eps = 1e-14
for i in range(10):
gc_iterations_param_dist = [10] * NUM_RUNS
epsilon_abs_param_dist = [eps] * NUM_RUNS
run_report_list = TestHarness.test_optimizer(
conjugate_gradients, test_func, x0_param,
[fprime for _ in range(NUM_RUNS)], gc_iterations_param_dist,
epsilon_abs_param_dist)
stats = StatsGenerator.generate_stats(run_report_list, (1, 1), 1)
print(stats)
csv_table.add_row([
gc_iterations_param_dist[i], epsilon_abs_param_dist[i],
CELL_FORMAT_STR.format(stats[0][0], stats[0][1]),
CELL_FORMAT_STR.format(stats[1][0], stats[1][1]),
CELL_FORMAT_STR.format(stats[2][0], stats[2][1])
])
eps *= 10
csv_table.export_table("function1_epsilon_variation.csv")
print("Finished function 1 epsilon variation")
csv_table.clear_table()
def problem_1_a():
csvf = CSVFormatter()
csvf.add_row(["Problem Size", "Time (ms)", "Profit"])
values = [
360, 83, 59, 130, 431, 67, 230, 52, 93, 125, 670, 892, 600, 38, 48,
147, 78, 256, 63, 17, 120, 164, 432, 35, 92, 110, 22, 42, 50, 323, 514,
28, 87, 73, 78, 15, 26, 78, 210, 36, 85, 189, 274, 43, 33, 10, 19, 389,
276, 312, 360, 83, 59, 130, 431, 67, 230, 52, 93, 125, 670, 892, 600,
38, 48, 147, 78, 256, 63, 17, 120, 164, 432, 35, 92, 110, 22, 42, 50,
323, 514, 28, 87, 73, 78, 15, 26, 78, 210, 36, 85, 189, 274, 43, 33,
10, 19, 389, 276, 312, 360, 83, 59, 130, 431, 67, 230, 52, 93, 125,
670, 892, 600, 38, 48, 147, 78, 256, 63, 17, 120, 164, 432, 35, 92,
110, 22, 42, 50, 323, 514, 28, 87, 73, 78, 15, 26, 78, 210, 36, 85,
189, 274, 43, 33, 10, 19, 389, 276, 312, 360, 83, 59, 130, 431, 67,
230, 52, 93, 125, 670, 892, 600, 38, 48, 147, 78, 256, 63, 17, 120,
164, 432, 35, 92, 110, 22, 42, 50, 323, 514, 28, 87, 73, 78, 15, 26,
78, 210, 36, 85, 189, 274, 43, 33, 10, 19, 389, 276, 312
]
weights = [
7, 0, 30, 22, 80, 94, 11, 81, 70, 64, 59, 18, 0, 36, 3, 8, 15, 42, 9,
0, 42, 47, 52, 32, 26, 48, 55, 6, 29, 84, 2, 4, 18, 56, 7, 29, 93, 44,
71, 3, 86, 66, 31, 65, 0, 79, 20, 65, 52, 13, 7, 0, 30, 22, 80, 94, 11,
81, 70, 64, 59, 18, 0, 36, 3, 8, 15, 42, 9, 0, 42, 47, 52, 32, 26, 48,
55, 6, 29, 84, 2, 4, 18, 56, 7, 29, 93, 44, 71, 3, 86, 66, 31, 65, 0,
79, 20, 65, 52, 13, 7, 0, 30, 22, 80, 94, 11, 81, 70, 64, 59, 18, 0,
36, 3, 8, 15, 42, 9, 0, 42, 47, 52, 32, 26, 48, 55, 6, 29, 84, 2, 4,
18, 56, 7, 29, 93, 44, 71, 3, 86, 66, 31, 65, 0, 79, 20, 65, 52, 13, 7,
0, 30, 22, 80, 94, 11, 81, 70, 64, 59, 18, 0, 36, 3, 8, 15, 42, 9, 0,
42, 47, 52, 32, 26, 48, 55, 6, 29, 84, 2, 4, 18, 56, 7, 29, 93, 44, 71,
3, 86, 66, 31, 65, 0, 79, 20, 65, 52, 13
]
num_runs = 5
knapsack_problem_size_range = (5, 125, 5)
max_weight_modifier = 37
harness_param_t0 = [25000] * num_runs
harness_param_t_final = [0.1] * num_runs
for problem_size in range(*knapsack_problem_size_range):
# create knapsack problem for test
kp = KnapsackProblem(max_weight=max_weight_modifier * problem_size,
values=values[:problem_size],
weights=weights[:problem_size])
harness_param_x0 = [[
round(random.uniform(0, 1)) for _ in range(problem_size)
] for _ in range(num_runs)]
allowed = [[0, 1] for _ in range(problem_size)]
harness_param_allowed = [allowed] * num_runs
run_report_list = TestHarness.test_optimizer(
simulated_annealing, kp.get_profit, harness_param_x0,
harness_param_allowed, harness_param_t0, harness_param_t_final)
time_stats = StatsGenerator.generate_stats(
list(map(lambda x: x[0], run_report_list)))
profit_stats = StatsGenerator.generate_stats(
list(map(lambda x: x[2], run_report_list)))
table_entry = [
problem_size,
StatsGenerator.stats_to_string(time_stats),
StatsGenerator.stats_to_string(profit_stats)
]
csvf.add_row(table_entry)
print("Finished problem size:", problem_size)
print(table_entry)
csvf.export_table("sa_knapsack_test.csv")