def test_schedule_random():
axis_map = AXIS_MAP.copy()
axis_map['variables'] = [
{
'name':
'timber',
'strategy':
'within_bounds', # Stay within range
'targets': (
[45.0] * 20, # 1 x time periods array of lower bounds
[55.0] * 20 # 1 x time periods array of upper bounds
),
'weight':
1.0
},
{
'name': 'carbon',
'strategy':
'cumulative_maximize', # target the max cumulative value
'weight': 1.0
},
{
'name': 'cost proxy',
'strategy':
'cumulative_minimize', # target the min cumulative value
'weight': 1.0
},
]
best, optimal_stand_rxs, vars_over_time = schedule(STAND_DATA,
axis_map,
VALID_MGMTS,
steps=15000,
report_interval=5000,
temp_min=0.006,
temp_max=20)
# stochastic process, just make sure it's in the ballpark
assert 5000 < best and 6000 > best
def test_schedule_random():
axis_map = AXIS_MAP.copy()
axis_map['variables'] = [
{
'name': 'timber',
'strategy': 'within_bounds', # Stay within range
'targets': (
[45.0] * 20, # 1 x time periods array of lower bounds
[55.0] * 20 # 1 x time periods array of upper bounds
),
'weight': 1.0
},
{
'name': 'carbon',
'strategy': 'cumulative_maximize', # target the max cumulative value
'weight': 1.0
},
{
'name': 'cost proxy',
'strategy': 'cumulative_minimize', # target the min cumulative value
'weight': 1.0
},
]
best, optimal_stand_rxs, vars_over_time = schedule(
STAND_DATA,
axis_map,
VALID_MGMTS,
steps=15000,
report_interval=5000,
temp_min=0.006,
temp_max=20
)
# stochastic process, just make sure it's in the ballpark
assert 5000 < best and 6000 > best
'weight':
1.0
},
{
'name': 'carbon',
'strategy':
'cumulative_maximize', # target the max cumulative value
'weight': 1.0
},
{
'name': 'cost proxy',
'strategy':
'cumulative_minimize', # target the min cumulative value
'weight': 1.0
},
]
#----------- STEP 3: Optimize (annealing over objective function) ---------#
best, optimal_stand_rxs, vars_over_time = schedule(stand_data,
axis_map,
valid_mgmts,
steps=250000,
report_interval=5000,
temp_min=0.0006,
temp_max=2,
live_plot=True)
#----------- STEP 4: output results ---------------------------------------#
print_results(axis_map, vars_over_time)
write_stand_mgmt_csv(optimal_stand_rxs, axis_map, filename="_results.csv")
'strategy': 'cumulative_minimize',
'weight': 1.0 },
]
#----------- STEP 3: Optimize (annealing over objective function) ---------#
# Do a quick run to get a good starting condition
best_start = float("inf")
for i in range(1):
best, optimal_stand_rxs, vars_over_time = schedule(
stand_data,
axis_map,
valid_mgmts,
steps=5000 + (1000*i),
report_interval=1000,
temp_min=1e-5,
temp_max=5.0
)
if best < best_start:
best_start = best
best_mgmts = optimal_stand_rxs
# Now run the full schedule
best, optimal_stand_rxs, vars_over_time = schedule(
stand_data,
axis_map,
valid_mgmts,
steps=295000,
report_interval=5000,
{
'name': 'cost proxy',
'strategy': 'cumulative_minimize',
'weight': 1.0
},
]
#----------- STEP 3: Optimize (annealing over objective function) ---------#
# Do a quick run to get a good starting condition
best_start = float("inf")
for i in range(1):
best, optimal_stand_rxs, vars_over_time = schedule(
stand_data,
axis_map,
valid_mgmts,
steps=5000 + (1000 * i),
report_interval=1000,
temp_min=1e-5,
temp_max=5.0)
if best < best_start:
best_start = best
best_mgmts = optimal_stand_rxs
# Now run the full schedule
best, optimal_stand_rxs, vars_over_time = schedule(
stand_data,
axis_map,
valid_mgmts,
steps=295000,
report_interval=5000,
temp_min=0.00005,
),
'weight': 1.0
},
{
'name': 'carbon',
'strategy': 'cumulative_maximize', # target the max cumulative value
'weight': 1.0
},
{
'name': 'cost proxy',
'strategy': 'cumulative_minimize', # target the min cumulative value
'weight': 1.0
},
]
#----------- STEP 3: Optimize (annealing over objective function) ---------#
best, optimal_stand_rxs, vars_over_time = schedule(
stand_data,
axis_map,
valid_mgmts,
steps=250000,
report_interval=5000,
temp_min=0.0006,
temp_max=2,
live_plot=True
)
#----------- STEP 4: output results ---------------------------------------#
print_results(axis_map, vars_over_time)
write_stand_mgmt_csv(optimal_stand_rxs, axis_map, filename="_results.csv")
