Freedom251 › Physics Simulation Software
EthicalSmartGrid
What EthicalSmartGrid is
EthicalSmartGrid is a free API package that simulates ethical choices in a simplified multi-agent smart grid. It provides a Gymnasium environment for reinforcement learning and documents a PettingZoo environment intended to work with PettingZoo-compliant algorithms with little or no modification. Households, offices, schools, or building parts act as agents representing inhabitants. They can consume and store energy from a shared local source, use personal storage, and buy from or sell energy to the national grid; scenarios can include prosumers that produce and consume energy. The use case addresses supply security, affordability, inclusiveness, and environmental sustainability. Observations combine shared global values with private agent values such as storage, comfort, and payoff. Numeric rewards are included, with optional argumentation-based rewards from version 1.2.0. Those rewards require the AJAR library. Users can configure agent numbers and profiles, energy constraints, and reward functions, and extend or replace components. The documentation says the simulator is simplified rather than realistic. The source code uses the MIT License.
Who it suits
EthicalSmartGrid suits researchers and developers building multi-agent reinforcement-learning scenarios around ethical energy decisions. It is intended for configurable experiments, not realistic grid simulation.
What it does well
- Free MIT-licensed source code
- Gymnasium and PettingZoo environments
- Configurable agents, profiles, and constraints
- Private local observations for each agent
- Numeric and optional argumentation rewards
What to know first
- Simulator is explicitly simplified
- Argumentation rewards require AJAR
- No lesson library or assessment tools
- Importing argumentation rewards without AJAR raises an error
Verdict
EthicalSmartGrid provides a configurable learning environment for studying ethical behavior in shared-energy scenarios. Its simplified model and AJAR dependency for argumentation rewards define the main boundaries for use.
Facts from EthicalSmartGrid's own pages
- Agents
- Agents represent buildings or parts of buildings, including households, offices, and schools, and act as proxies for their inhabitants. · ethicsai.github.io, 11 Oct 2026
- Assessment tools
- No · github.com, 24 Sept 2026
- Authoring tools
- Yes · github.com, 24 Sept 2026
- Customization
- Users can configure scenarios with different agent numbers and profiles, energy constraints, and reward functions, and can extend or replace components. · ethicsai.github.io, 11 Oct 2026
- Custom rewards
- Users can define custom argumentation-based reward functions with the AJAR library. · ethicsai.github.io, 11 Oct 2026
- Energy model
- Agents can consume and store energy from a shared local source, use personal storage, and buy from or sell energy to the national grid. · ethicsai.github.io, 11 Oct 2026
- Ethical values
- Its use case considers security of supply, affordability, inclusiveness, and environmental sustainability. · ethicsai.github.io, 11 Oct 2026
- Framework
- It provides a Gymnasium environment for multi-agent reinforcement learning and aims to follow the Gymnasium API closely. · ethicsai.github.io, 11 Oct 2026
- Framework compatibility
- The documented usage provides a PettingZoo environment designed to work with PettingZoo-compliant learning algorithms with little or no modification. · ethicsai.github.io, 11 Oct 2026
- Free plan
- Yes · github.com, 24 Sept 2026
- Installation
- The documentation says to install it with `pip install ethical-smart-grid` or clone the project repository. · ethicsai.github.io, 11 Oct 2026
- Lesson library
- No · github.com, 24 Sept 2026
- License
- The repository says the source code is licensed under the MIT License and notes that some included data may have other licenses. · github.com, 11 Oct 2026
- Model limitation
- The project documentation says its simulated smart grid is simplified and is not developed as a realistic simulator. · ethicsai.github.io, 11 Oct 2026
- Observations
- Observations are split into shared global values and agent-specific local values, including personal storage, comfort, and payoff. · ethicsai.github.io, 11 Oct 2026
- Optional dependency
- Using the argumentation reward package requires the AJAR library; importing that package without AJAR raises an error. · ethicsai.github.io, 11 Oct 2026
- Privacy
- The use-case separates shared global observations from individual local observations, which other agents cannot access. · ethicsai.github.io, 11 Oct 2026
- Purpose
- EthicalSmartGrid is a simplified multi-agent smart-grid simulator for learning ethical behaviours. · ethicsai.github.io, 11 Oct 2026
- Reward functions
- The package includes numeric reward functions, and from version 1.2.0 it also supports argumentation-based reward functions. · ethicsai.github.io, 11 Oct 2026
- Rewards
- It provides numeric reward functions and, from v1.2.0, optional argumentation-based reward functions. · ethicsai.github.io, 11 Oct 2026
- Simulation limitation
- The documentation says the simulated grid is simplified and is not developed as a realistic simulator. · ethicsai.github.io, 11 Oct 2026
- Support
- The project README directs questions to GitHub Discussions and bug reports to the issue tracker. · github.com, 11 Oct 2026
- 3D visuals
- No · github.com, 24 Sept 2026
- Use case
- The simulated grid includes prosumer agents that produce and consume energy, a shared local energy source, and a connection to the national grid. · ethicsai.github.io, 11 Oct 2026
Where EthicalSmartGrid stands on Freedom251
- Physics Simulation Software#81 of 89