Simulation data and code for energy-neutral evaluation and compact reinforcement learning control of heterogeneous islanded microgrids

Description

Preserved simulation and analysis evidence supporting the associated Energy manuscript. Includes 420 distinct saved evaluation traces across main, carbon/attention, low-SOC and compact-policy evaluations; completion flags for every attempt; selected actor checkpoints; original training histories; processed numerical evidence; simulation inputs; analysis and reproduction code; model artifacts; variable dictionaries; figure/table mappings; provenance and SHA-256 checksums. The package preserves failed attempts and endpoint-specific pairing. It supports offline verification without MATLAB; full retraining requires the documented MATLAB/Simulink and Python dependencies. Hydroturbine inputs are found typical values used for simulation, not author-collected measurements. Download all 70 Energy_data_partNN.zip files and extract them into the same directory, preserving relative paths; see DATASET_README.md and parts_manifest.json. The partitioning changes only archive grouping, not research files. Contributors are affiliated with the School of Electrical Engineering, Xi'an Jiaotong University, Xi'an 710049, China. Keywords: islanded microgrid; frequency regulation; soft actor-critic; terminal-energy recovery; structured pruning; simulation data.

Steps to reproduce

Extract the ZIP archive and read README.md and 复现工程/REPRODUCTION.md. From 复现工程, run python check_key_results.py for Python-only checks of archived results. Run python ../verify_package.py to check file integrity. Optional offline table recomputation is available through recompute_main_tables.py. Full controller retraining and evaluation require MATLAB/Simulink, a compatible MATLAB Engine for Python, the model toolboxes and the documented Python dependencies. The archived data include incomplete evaluation attempts; use the recorded completion flags and endpoint-specific valid pairs.

Institutions

• Xi'an Jiaotong University

  Shaanxi, Xi'an

Categories

Licence