Overview

Overview#

eprllib was born out of the need to bridge the gap between building modeling with EnergyPlus and Reinforcement Learning (RL). Traditionally, integrating these two disciplines has been complex and laborious. eprllib aims to simplify this process, offering an intuitive and flexible interface for developing intelligent agents that interact with building simulations.

Purpose and Scope#

The main objective of eprllib is to facilitate experimentation with Reinforcement Learning algorithms in the context of energy efficiency and building control. The library allows researchers and professionals to design, train, and evaluate agents that can optimize energy consumption, improve occupant comfort, and manage resources more efficiently.

Key Features#

eprllib stands out for the following features:

  • Native Multi-Agent Environment: eprllib is designed to work with complex multi-agent systems, including hierarchical architectures. This allows for modeling scenarios where multiple agents interact with each other and the building environment. It also supports the use of individual agents for simpler cases.

  • Flexibility and Configurability: The library offers configurable modules for Filters , Triggers, and Rewards.

    This allows users to customize the learning environment to suit their specific needs and explore different control strategies.

  • Inter-Agent Communication and Coordination: eprllib includes a dedicated module for communication and coordination between agents. This is essential for implementing collaborative strategies and solving complex problems that require the interaction of multiple agents.

  • Deep Integration with EnergyPlus: Thanks to its close integration with the EnergyPlus API, eprllib allows access to all the variables, metrics, parameters, and actuators that EnergyPlus offers. This means that users can work with detailed and realistic building models, covering a wide range of case studies, from thermostat control to the management of complete HVAC systems.

  • Ease of Use: eprllib has been developed following a scheme similar to RLlib, a popular Reinforcement Learning library. This facilitates the learning of eprllib for those users familiar with RLlib.

Target Audience and Prior Knowledge#

eprllib is primarily aimed at researchers with experience in EnergyPlus and an interest in Reinforcement Learning. It is assumed that users have prior knowledge in both areas, as eprllib has been developed following a scheme similar to RLlib to facilitate the joint learning of both tools. While it is not necessary to be an expert in RL, a basic understanding of fundamental concepts such as agents, environments, rewards, and policies is recommended.

Software and Hardware Requirements#

eprllib requires Python 3.10 or higher and EnergyPlus 9.3 or higher, with version 24.2 of the latter being recommended. In addition, RLlib’s specific dependencies will be required, which will be installed automatically during the installation of eprllib. Hardware requirements will depend on the complexity of the building models and the RL algorithms used. It is recommended to have a computer with sufficient RAM and processing capacity to run complex simulations.

Installation#

Installing eprllib is simple and is done through pip:

pip install eprllib

This command will install eprllib and all its dependencies, including RLlib and the necessary libraries for communication with EnergyPlus.

Example of Use#

[Here, a basic example of use would be included to illustrate how to use eprllib to implement an agent in a simulated EnergyPlus environment. This example will be completed later.]

  1. Import eprllib.

  2. Configure an EnvConfig object to feed EnergyPlus Python API and RLlib with the environment configuration based on the EnergyPlus model,

    specifying the parameters required (see eprllib.Env.EnvConfig).

  3. Configure RLlib algorithm to train the policy.

  4. Execute the training using RLlib or Tune.

License and Additional Resources#

eprllib is distributed under the MIT license, a permissive open source license that allows its use and modification for both personal and commercial purposes.

For more information, to report issues, or to contribute to the development of eprllib, you can visit the GitHub repository. There you will also find a discussion forum for each version (Release) where you can interact with other users and developers.

Contents