Assume-framework

**What has changed:**
- The learning performance has been improved
- The learning can now be also performed on CUDA enabled devices
- Added tracking of evaluation periods for better learning performance evaluation
- Storage units behavior bugs have been adressed
- Now several simulation can be started in parallel
- New Grafana dashboard definitions for easier analysis
- Docker compose file now includes Renderer to save plots directly from Grafan dashboards

This is the initial release of assume-framework, which is published to PyPi.

The assume-framework allows to define different energy market designs and includes reinforcement learning capabilities.


market_paper
This tag is used to release a citable version of the code used for the conference paper:
"Market Abstraction of Energy Markets and Policies – Application in an Agent-Based Modeling Toolbox"