Looking for a way to map observations to the schema

NeurIPS 2022: CityLearn Challenge
#1

As a baseline I want to expand the simple rule based agent with some more simple rules. To create sensible features for this it would really help to be able to map the observations in the agent to their names. My first effort to do this was to use the schema.json file, but it does not really seem to line up. For instance the Solar Generation seems to be missing.

In short my questions is there an easy way to map observation values to their names?

#2

So far came to to the following mapping. Still some blanks, but I’ll update this list when I they become clear.

column_name = [  'M',
    'D',
    'H',
    'T',
    'T_6h',
    'T_12h',
    'T_24h',
    'Hum', 
    'Hum_6h', 
    'Hum_12h',    
    'Hum_24h', 
    'dhi', 
    'dhi_6h', 
    'dhi_12h', 
    'dhi_24h',
    'dni', 
    'dni_6h',
    'dni_12h', 
    'dni_24h',
    'carbon?',
    'consumption',
    'generation?',  
    'unkown',
    'generation?',
    'price', 
    'price_6h',
    'price_12h', 
    'price_24h',
]
#3

I had the same issue and made a similar assumption. Here is my take.

{‘month’: True,
‘day_type’: True,
‘hour’: True,
‘outdoor_dry_bulb_temperature’: True,
‘outdoor_dry_bulb_temperature_predicted_6h’: True,
‘outdoor_dry_bulb_temperature_predicted_12h’: True,
‘outdoor_dry_bulb_temperature_predicted_24h’: True,
‘outdoor_relative_humidity’: True,
‘outdoor_relative_humidity_predicted_6h’: True,
‘outdoor_relative_humidity_predicted_12h’: True,
‘outdoor_relative_humidity_predicted_24h’: True,
‘diffuse_solar_irradiance’: True,
‘diffuse_solar_irradiance_predicted_6h’: True,
‘diffuse_solar_irradiance_predicted_12h’: True,
‘diffuse_solar_irradiance_predicted_24h’: True,
‘direct_solar_irradiance’: True,
‘direct_solar_irradiance_predicted_6h’: True,
‘direct_solar_irradiance_predicted_12h’: True,
‘direct_solar_irradiance_predicted_24h’: True,
‘carbon_intensity’: True,
‘non_shiftable_load’: True,
‘solar_generation’: True,
‘electrical_storage_soc’: True,
‘net_electricity_consumption’: True,
‘electricity_pricing’: True,
‘electricity_pricing_predicted_6h’: True,
‘electricity_pricing_predicted_12h’: True,
‘electricity_pricing_predicted_24h’: True}

This in order of states with value true in the schema.json file. I generated this while preparing a state_action_schema json file for MARLISA. I also don’t know if this is correct but I know the month and day_type index are. Though i know this, I would also like some clarifications/corrections/.

#4

@eric_leerssen thanks for your question. If you want to know get the name list of active observations in each building, use env.buildings[0].active_observations or env.buildings[0].observations for a dictionary of current time step observations. The ordering of the observations follows the ordering in the schema and excludes observations where active:false in schema.

The documentation of available observations can be found in the docs.

Hope these help?

1 Like