GreenOpt-AI v2.4.3: Reproducibility Package for Validation-First Reinforcement Learning for Resource-Efficient Greenhouse Control
Year: 2026
Type: article
Abstract: GreenOpt-AI v2.4.3 is the publication reproducibility archive for a validation-first reinforcement-learning framework for resource-efficient greenhouse control. The release contains the final source implementation, structured-action and empirical-support mechanisms, V2.4.3 validation evidence, final TD3/DDPG/SAC evaluation results across five independent training seeds and 16 paired eight-week evaluation windows, a matched historical Reference control-schedule replay, statistical-analysis tables, manuscript figure-generation code, provenance records, and cryptographic checksums. The archive supports exact reproduction of the reported publication analyses and figures from the frozen evaluation evidence. The study is simulation-based; reported crop and resource effects are simulator-derived and are not presented as physical-greenhouse field validation.
Author: Hindi, Iman
Institutions: University of Jordan, Al Hussein Technical University
Language: English
Subfield: Plant Science
Domain: Life Sciences
SDG: Climate action
Open Access status: green
OpenAlex ID: W7224939607