-
Notifications
You must be signed in to change notification settings - Fork 1
/
config.py
492 lines (483 loc) · 23.1 KB
/
config.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
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
config_dict = {
# House properties
# The house is modelled as a 10mx10m square with 2.5m height, 8 windows of 1.125 m² and 2 doors of 2m² each.
# The formulas for Ua, Cm, Ca and Hm are mainly taken here: http://gridlab-d.shoutwiki.com/wiki/Residential_module_user's_guide
"default_house_prop": {
"id": 1,
"init_air_temp": 20,
"init_mass_temp": 20,
"target_temp": 20,
"deadband": 0,
"Ua": 2.18e02, # House walls conductance (W/K). Multiplied by 3 to account for drafts (according to https://dothemath.ucsd.edu/2012/11/this-thermal-house/)
"Cm": 3.45e06, # House thermal mass (J/K) (area heat capacity:: 40700 J/K/m2 * area 100 m2)
"Ca": 9.08e05, # Air thermal mass in the house (J/K): 3 * (volumetric heat capacity: 1200 J/m3/K, default area 100 m2, default height 2.5 m)
"Hm": 2.84e03, # House mass surface conductance (W/K) (interioor surface heat tansfer coefficient: 8.14 W/K/m2; wall areas = Afloor + Aceiling + Aoutwalls + Ainwalls = A + A + (1+IWR)*h*R*sqrt(A/R) = 455m2 where R = width/depth of the house (default R: 1.5) and IWR is I/O wall surface ratio (default IWR: 1.5))
"window_area": 7.175, # Gross window area, in m^2
"shading_coeff": 0.67, # Window Solar Heat Gain Coefficient, look-up table in Gridlab reference
"solar_gain_bool": True, # Boolean to model the solar gain
},
"noise_house_prop": {
"noise_mode": "big_start_temp", # Can be: no_noise, small_noise, big_noise, small_start_temp, big_start_temp
"noise_parameters": {
"no_noise": {
"std_start_temp": 0, # Std noise on starting temperature
"std_target_temp": 0, # Std Noise on target temperature
"factor_thermo_low": 1, # Lowest random factor for Ua, Cm, Ca, Hm
"factor_thermo_high": 1, # Highest random factor for Ua, Cm, Ca, Hm
},
"dwarf_noise": {
"std_start_temp": 0.05, # Std noise on starting temperature
"std_target_temp": 0.05, # Std Noise on target temperature
"factor_thermo_low": 1, # Lowest random factor for Ua, Cm, Ca, Hm
"factor_thermo_high": 1, # Highest random factor for Ua, Cm, Ca, Hm
},
"house_small_noise": {
"std_start_temp": 0, # Std noise on starting temperature
"std_target_temp": 0, # Std Noise on target temperature
"factor_thermo_low": 0.9, # Lowest random factor for Ua, Cm, Ca, Hm
"factor_thermo_high": 1.1, # Highest random factor for Ua, Cm, Ca, Hm
},
"house_medium_noise": {
"std_start_temp": 0, # Std noise on starting temperature
"std_target_temp": 0, # Std Noise on target temperature
"factor_thermo_low": 0.8, # Lowest random factor for Ua, Cm, Ca, Hm
"factor_thermo_high": 1.2, # Highest random factor for Ua, Cm, Ca, Hm
},
"house_big_noise": {
"std_start_temp": 0, # Std noise on starting temperature
"std_target_temp": 0, # Std Noise on target temperature
"factor_thermo_low": 0.5, # Lowest random factor for Ua, Cm, Ca, Hm
"factor_thermo_high": 1.5, # Highest random factor for Ua, Cm, Ca, Hm
},
"small_noise": {
"std_start_temp": 3, # Std noise on starting temperature
"std_target_temp": 1, # Std Noise on target temperature
"factor_thermo_low": 0.9, # Lowest random factor for Ua, Cm, Ca, Hm
"factor_thermo_high": 1.1, # Highest random factor for Ua, Cm, Ca, Hm
},
"big_noise": {
"std_start_temp": 5, # Std noise on starting temperature
"std_target_temp": 2, # Std Noise on target temperature
"factor_thermo_low": 0.8, # Lowest random factor for Ua, Cm, Ca, Hm
"factor_thermo_high": 1.2, # Highest random factor for Ua, Cm, Ca, Hm
},
"small_start_temp": {
"std_start_temp": 3, # Std noise on starting temperature
"std_target_temp": 0, # Std Noise on target temperature
"factor_thermo_low": 1, # Lowest random factor for Ua, Cm, Ca, Hm
"factor_thermo_high": 1, # Highest random factor for Ua, Cm, Ca, Hm
},
"big_start_temp": {
"std_start_temp": 5, # Std noise on starting temperature
"std_target_temp": 0, # Std Noise on target temperature
"factor_thermo_low": 1, # Lowest random factor for Ua, Cm, Ca, Hm
"factor_thermo_high": 1, # Highest random factor for Ua, Cm, Ca, Hm
},
},
},
"noise_house_prop_test": {
"noise_mode": "small_start_temp", # Can be: no_noise, small_noise, big_noise, small_start_temp, big_start_temp
"noise_parameters": {
"no_noise": {
"std_start_temp": 0, # Std noise on starting temperature
"std_target_temp": 0, # Std Noise on target temperature
"factor_thermo_low": 1, # Lowest random factor for Ua, Cm, Ca, Hm
"factor_thermo_high": 1, # Highest random factor for Ua, Cm, Ca, Hm
},
"dwarf_noise": {
"std_start_temp": 0.05, # Std noise on starting temperature
"std_target_temp": 0.05, # Std Noise on target temperature
"factor_thermo_low": 1, # Lowest random factor for Ua, Cm, Ca, Hm
"factor_thermo_high": 1, # Highest random factor for Ua, Cm, Ca, Hm
},
"small_noise": {
"std_start_temp": 3, # Std noise on starting temperature
"std_target_temp": 1, # Std Noise on target temperature
"factor_thermo_low": 0.9, # Lowest random factor for Ua, Cm, Ca, Hm
"factor_thermo_high": 1.1, # Highest random factor for Ua, Cm, Ca, Hm
},
"big_noise": {
"std_start_temp": 5, # Std noise on starting temperature
"std_target_temp": 2, # Std Noise on target temperature
"factor_thermo_low": 0.8, # Lowest random factor for Ua, Cm, Ca, Hm
"factor_thermo_high": 1.2, # Highest random factor for Ua, Cm, Ca, Hm
},
"small_start_temp": {
"std_start_temp": 3, # Std noise on starting temperature
"std_target_temp": 0, # Std Noise on target temperature
"factor_thermo_low": 1, # Lowest random factor for Ua, Cm, Ca, Hm
"factor_thermo_high": 1, # Highest random factor for Ua, Cm, Ca, Hm
},
"big_start_temp": {
"std_start_temp": 5, # Std noise on starting temperature
"std_target_temp": 0, # Std Noise on target temperature
"factor_thermo_low": 1, # Lowest random factor for Ua, Cm, Ca, Hm
"factor_thermo_high": 1, # Highest random factor for Ua, Cm, Ca, Hm
},
},
},
# HVAC properties
"default_hvac_prop": {
"id": 1,
"COP": 2.5, # Coefficient of performance (power spent vs heat displaced)
"cooling_capacity": 15000, # Cooling capacity (W)
"latent_cooling_fraction": 0.35, # Fraction of latent cooling w.r.t. sensible cooling
"lockout_duration": 40, # In seconds
"lockout_noise": 0, # In seconds
},
"noise_hvac_prop": {
"noise_mode": "no_noise", # Can be: no_noise, small_noise, big_noise
"noise_parameters": {
"no_noise": {
"cooling_capacity_list": {10000: [10000], 15000: [15000]}
# "std_latent_cooling_fraction": 0, # Std Gaussian noise on latent_cooling_fraction
# "factor_COP_low": 1, # Lowest random factor for COP
# "factor_COP_high": 1, # Highest random factor for COP
# "factor_cooling_capacity_low": 1, # Lowest random factor for cooling_capacity
# "factor_cooling_capacity_high": 1, # Highest random factor for cooling_capacity
},
"small_noise": {
"cooling_capacity_list": {10000: [9000, 10000, 11000], 15000: [12500, 15000, 17500]}
# "std_latent_cooling_fraction": 0.05, # Std Gaussian noise on latent_cooling_fraction
# "factor_COP_low": 0.95, # Lowest random factor for COP
# "factor_COP_high": 1.05, # Highest random factor for COP
# "factor_cooling_capacity_low": 0.9, # Lowest random factor for cooling_capacity
# "factor_cooling_capacity_high": 1.1, # Highest random factor for cooling_capacity
},
"big_noise": {
"cooling_capacity_list": {
10000: [7500, 9000, 10000, 11000, 12500],
15000: [10000, 12500, 15000, 17500, 20000],
}
# "std_latent_cooling_fraction": 0.1, # Std Gaussian noise on latent_cooling_fraction
# "factor_COP_low": 0.85, # Lowest random factor for COP
# "factor_COP_high": 1.15, # Highest random factor for COP
# "factor_cooling_capacity_low": 0.6666667, # Lowest random factor for cooling_capacity
# "factor_cooling_capacity_high": 1.3333333333, # Highest random factor for cooling_capacity
},
},
},
"noise_hvac_prop_test": {
"noise_mode": "no_noise", # Can be: no_noise, small_noise, big_noise
"noise_parameters": {
"no_noise": {
"std_latent_cooling_fraction": 0, # Std Gaussian noise on latent_cooling_fraction
"factor_COP_low": 1, # Lowest random factor for COP
"factor_COP_high": 1, # Highest random factor for COP
"factor_cooling_capacity_low": 1, # Lowest random factor for cooling_capacity
"factor_cooling_capacity_high": 1, # Highest random factor for cooling_capacity
},
"small_noise": {
"std_latent_cooling_fraction": 0.05, # Std Gaussian noise on latent_cooling_fraction
"factor_COP_low": 0.95, # Lowest random factor for COP
"factor_COP_high": 1.05, # Highest random factor for COP
"factor_cooling_capacity_low": 0.9, # Lowest random factor for cooling_capacity
"factor_cooling_capacity_high": 1.1, # Highest random factor for cooling_capacity
},
"big_noise": {
"std_latent_cooling_fraction": 0.1, # Std Gaussian noise on latent_cooling_fraction
"factor_COP_low": 0.85, # Lowest random factor for COP
"factor_COP_high": 1.15, # Highest random factor for COP
"factor_cooling_capacity_low": 0.6666667, # Lowest random factor for cooling_capacity
"factor_cooling_capacity_high": 1.3333333333, # Highest random factor for cooling_capacity
},
},
},
# Env properties
"default_env_prop": {
"start_datetime": "2021-01-01 00:00:00", # Start date and time (Y-m-d H:M:S)
"start_datetime_mode": "random", # Can be random (randomly chosen in the year after original start_datetime) or fixed (stays as the original start_datetime)
"time_step": 4, # Time step in seconds
"cluster_prop": {
"temp_mode": "noisy_sinusoidal_heatwave", # Can be: constant, sinusoidal, noisy_sinusoidal
"temp_parameters": {
"constant": {
"day_temp": 26.5, # Day temperature
"night_temp": 26.5, # Night temperature
"temp_std": 0, # Noise std dev on the temperature
"random_phase_offset": False,
},
"sinusoidal": {
"day_temp": 30,
"night_temp": 23,
"temp_std": 0,
"random_phase_offset": False,
},
"sinusoidal_hot": {
"day_temp": 30,
"night_temp": 28,
"temp_std": 0,
"random_phase_offset": False,
},
"sinusoidal_heatwave": {
"day_temp": 34,
"night_temp": 28,
"temp_std": 0,
"random_phase_offset": False,
},
"sinusoidal_hot_heatwave": {
"day_temp": 38,
"night_temp": 32,
"temp_std": 0,
"random_phase_offset": False,
},
"sinusoidal_cold_heatwave": {
"day_temp": 30,
"night_temp": 24,
"temp_std": 0,
"random_phase_offset": False,
},
"sinusoidal_cold": {
"day_temp": 24,
"night_temp": 22,
"temp_std": 0,
"random_phase_offset": False,
},
"noisy_sinusoidal": {
"day_temp": 30,
"night_temp": 23,
"temp_std": 0.5,
"random_phase_offset": False,
},
"noisy_sinusoidal_hot": {
"day_temp": 30,
"night_temp": 28,
"temp_std": 0.5,
"random_phase_offset": False,
},
"noisy_sinusoidal_heatwave": {
"day_temp": 34,
"night_temp": 28,
"temp_std": 0.5,
"random_phase_offset": False,
},
"noisier_sinusoidal_heatwave": {
"day_temp": 34,
"night_temp": 28,
"temp_std": 2,
"random_phase_offset": False,
},
"noisy_sinusoidal_cold": {
"day_temp": 24,
"night_temp": 22,
"temp_std": 0.5,
"random_phase_offset": False,
},
"shifting_sinusoidal": {
"day_temp": 30,
"night_temp": 23,
"temp_std": 0,
"random_phase_offset": True,
},
"shifting_sinusoidal_heatwave": {
"day_temp": 34,
"night_temp": 28,
"temp_std": 0,
"random_phase_offset": True,
},
},
"nb_agents": 1, # Number of houses
"nb_agents_comm": 10, # Maximal number of houses a single house communicates with
"agents_comm_mode": "neighbours", # Communication mode
"comm_defect_prob": 0, # Probability of a communication link being broken
"agents_comm_parameters": {
"neighbours_2D": {
"row_size": 5, # Row side length
"distance_comm": 2, # Max distance between two communicating houses
},
},
},
"state_properties": {
"hour": False,
"day": False,
"solar_gain": False,
"thermal": False,
"hvac": False,
},
"message_properties": {
"thermal": False,
"hvac": False,
},
"power_grid_prop": {
"base_power_mode": "interpolation", # Interpolation (based on deadband bang-bang controller) or constant
"base_power_parameters": {
"constant": {
"avg_power_per_hvac": 4200, # Per hvac. In Watts.
"init_signal_per_hvac": 910, # Per hvac.
},
"interpolation": {
"path_datafile": "./monteCarlo/mergedGridSearchResultFinal.npy",
"path_parameter_dict": "./monteCarlo/interp_parameters_dict.json",
"path_dict_keys": "./monteCarlo/interp_dict_keys.csv",
"interp_update_period": 300, # Seconds
"interp_nb_agents": 100, # Max number of agents over which the interpolation is run
},
},
"artificial_signal_ratio_range": 1, # Scale of artificial multiplicative factor randomly multiplied (or divided) at each episode during training. Ex: 1 will not modify signal. 3 will have signal between 33% and 300% of what is computed.
"artificial_ratio": 1.0,
"signal_mode": "perlin", # Mode of the signal. Currently available: flat, sinusoidal, regular_steps, perlin
"signal_parameters": {
"flat": {},
"sinusoidals": {
"periods": [400, 1200], # In seconds
"amplitude_ratios": [0.1, 0.3], # As a ratio of avg_power_per_hvac
},
"regular_steps": {
"amplitude_per_hvac": 6000, # In watts
"period": 300, # In seconds
},
"perlin": {
"amplitude_ratios": 0.9,
"nb_octaves": 5,
"octaves_step": 5,
"period": 400,
},
"amplitude+_perlin": {
"amplitude_ratios": 0.9*1.1,
"nb_octaves": 5,
"octaves_step": 5,
"period": 400,
},
"amplitude++_perlin": {
"amplitude_ratios": 0.9*1.3,
"nb_octaves": 5,
"octaves_step": 5,
"period": 400,
},
"fast+_perlin": {
"amplitude_ratios": 0.9,
"nb_octaves": 5,
"octaves_step": 5,
"period": 300,
},
"fast++_perlin": {
"amplitude_ratios": 0.9,
"nb_octaves": 5,
"octaves_step": 5,
"period": 200,
},
},
},
"reward_prop": {
"alpha_temp": 1, # Tradeoff parameter for temperature in the loss function: alpha_temp * temperature penalty + alpha_sig * regulation signal penalty.
"alpha_sig": 1, # Tradeoff parameter for signal in the loss function: alpha_temp * temperature penalty + alpha_sig * regulation signal penalty.
"norm_reg_sig": 7500, # Average power use, for signal normalization
"temp_penalty_mode": "individual_L2", # Mode of temperature penalty
"temp_penalty_parameters": {
"individual_L2": {},
"common_L2": {},
"common_max_error": {},
"mixture": {
"alpha_ind_L2": 1,
"alpha_common_L2": 1,
"alpha_common_max": 0,
},
},
"sig_penalty_mode": "common_L2", # Mode of signal penalty
},
},
# Agent properties
"PPO_prop": {
"actor_layers": [100, 100],
"critic_layers": [100, 100],
"gamma": 0.99,
"lr_critic": 3e-3,
"lr_actor": 1e-3,
"clip_param": 0.2,
"max_grad_norm": 0.5,
"ppo_update_time": 10,
"batch_size": 256,
"zero_eoepisode_return": False,
},
"MAPPO_prop": {
"actor_layers": [100, 100],
"critic_layers": [100, 100],
"gamma": 0.99,
"lr_critic": 3e-3,
"lr_actor": 1e-3,
"clip_param": 0.2,
"max_grad_norm": 0.5,
"ppo_update_time": 10,
"batch_size": 256,
"zero_eoepisode_return": False,
},
"DDPG_prop": {
"actor_hidden_dim": 256,
"critic_hidden_dim": 256,
"gamma": 0.99,
"lr_critic": 3e-3,
"lr_actor": 1e-3,
"soft_tau": 0.01,
"clip_param": 0.2,
"max_grad_norm": 0.5,
"ddpg_update_time": 10,
"batch_size": 64,
"buffer_capacity": 524288,
"episode_num": 10000,
"learn_interval": 100,
"learn_interval": 100,
"random_steps": 100,
"gumbel_softmax_tau": 1,
"DDPG_shared": True
},
"TarMAC_prop": {
"recurrent_policy": True, # Use RNN
"state_size": 128, # Size of the RNN state
"communication_size": 32, # Size of the communication message
"tarmac_communication_mode": "from_states_rec_att", # Mode of communication protocole (not the same as communication_mode of the cluster)
"comm_num_hops": 1, # Number of hops during the communication
"value_loss_coef": 0.5, # Coefficient of the value loss in the loss function
"entropy_coef": 0.01, # Coefficient of the entropy loss in the loss function
"tarmac_lr": 7e-4, # Learning rate
"tarmac_eps": 1e-5, # Epsilon for RMSProp or Adam optimizer
"tarmac_gamma": 0.99, # Discount factor
"tarmac_alpha": 0.99, # Alpha for RMSProp optimizer
"tarmac_max_grad_norm": 0.5, # Maximal norm of the gradient. If None, no clipping is done.
"distributed": False,
"nb_tarmac_updates": 10,
"tarmac_batch_size": 128,
},
"TarMAC_PPO_prop": {
"actor_hidden_state_size": 64, # Size of the hidden state of the actor
"critic_hidden_layer_size": 64, # Size of the hidden layers in the critic
"communication_size": 16, # Size of the communication message
"key_size": 8, # Size of the key/query
"comm_num_hops": 1, # Number of hops during the communication
"lr_critic": 1e-3,
"lr_actor": 1e-3,
"eps": 1e-5, # Epsilon for RMSProp or Adam optimizer
"gamma": 0.99, # Discount factor
"max_grad_norm": 0.5, # Maximal norm of the gradient. If None, no clipping is done.
"clip_param": 0.2,
"ppo_update_time": 10,
"batch_size": 256,
"with_gru": False,
"with_comm": True,
"number_agents_comm_tarmac": 10,
"tarmac_comm_mode": 'neighbours',
"tarmac_comm_defect_prob": 0.0 # Probability of a TarMAC communication defect
},
"DQN_prop": {
"network_layers": [100, 100],
"gamma": 0.99,
"tau": 0.01,
"buffer_capacity": 524288,
"lr": 1e-3,
"batch_size": 256,
"epsilon_decay": 0.99998,
"min_epsilon": 0.01,
},
"MPC_prop": {
"rolling_horizon": 15,
},
# Training process properties
"training_prop": {
"nb_inter_saving_actor": 9, # Number of intermediate saving of the actor
"nb_test_logs": 200, # Number of times the agent is tested during the training
"nb_time_steps_test": 21600, # Number of time steps during the tests (1 full day)
"nb_tr_episodes": 200, # Number of training episodes (environment resets)
"nb_tr_epochs": 200, # Number of training epochs (policy updates)
"nb_tr_logs": 200, # Number of times the performances are averaged and logged during the training
"nb_time_steps": 3276800, # Number of time steps during the training
},
}