Skip to content

Latest commit

 

History

History
41 lines (36 loc) · 2.26 KB

3.1_predefined_envs.md

File metadata and controls

41 lines (36 loc) · 2.26 KB

Here, we give an overview of the pre-registered environments and their naming convention

N fertilization environments

The fertilization environments differ along the following dimensions: weather generation, location, duration. All these environments are readily available and can be created with

import gym
gym.make(id=env_id)

where env_id=f{'Corn{duration}{location}{weather}-v1'}. Below, we explain the values that these variables can have and their corresponding meaning.

  • duration indicates time horizon of the experiment. It can take the values Short (1 year), Middle (2 years), or Long (5 years).
  • location indicates the location where the experiment takes place, which affects the historical weather data that is used. It can take the values RockSprings or NewHolland.
  • weather indicates how the weather is generated, which can either be random (random shuffled years from historical
  • data) or fixed. It can take the values RW (random) or FW (fixed).

All of these environments come with costs that can be used to define constraints on the number of fertilization events, the amount of N applied, and N leaching, volatilization, and emission. Following the interface from safety gym, these costs are given in the info dictionary that is returned by the step function of each environment, which does not break the standard OpenAI gym interface.

Crop planning environments

The crop planning environments differ along the following dimensions: weather generation, location. The duration is fixed to 19 years as these experiments only make sense over long time horizons. All these environments are readily available and can be created with

import gym
gym.make(id=env_id)

where env_id=f{'CropPlanning{location}{weather}-v1'}. Below, we explain the values that these variables can have and their corresponding meaning.

  • location indicates the location where the experiment takes place, which affects the historical weather data that is used. It can take the values RockSprings or NewHolland.
  • weather indicates how the weather is generated, which can either be random (random shuffled years from historical data) or fixed. It can take the values RW (random) or FW (fixed).