Calibration-First Reward-Component Auditing for Reinforcement Learning Control in Smart Greenhouses

智能温室强化学习控制中基于校准优先的奖励组件审计

Abstract: Greenhouse reinforcement learning can test climate-control ideas at a speed and scale that is difficult to achieve with crop experiments alone. For smart-greenhouse control, however, a single simulator return is not enough: a grower or control engineer also needs to know when the policy heats, enriches CO2, vents, manages humidity, deploys screens, or uses energy.

摘要： 温室强化学习能够以仅靠作物实验难以实现的规模和速度测试气候控制方案。然而，对于智能温室控制而言，单一的模拟器回报（return）是不够的：种植者或控制工程师还需要了解策略在何时进行加热、CO2 补给、通风、湿度管理、遮阳幕布部署或能源消耗。

We propose a reproducible calibration-first reward audit framework that keeps named greenhouse-control reward components comparable across simulator training, facility-adapted rollouts, logged Autonomous Greenhouse Challenge records, and actuator-rule distillation.

我们提出了一种可复现的“校准优先”奖励审计框架，该框架确保了命名的温室控制奖励组件在模拟器训练、设施适配部署、自动温室挑战赛（Autonomous Greenhouse Challenge）记录以及执行器规则蒸馏过程中保持可比性。

In GreenLight-Gym, the framework decomposes the scalar reward into conditional temperature, CO2, humidity and vapor-pressure-deficit, screen, and actuation-proxy terms; adapts GreenLight to the second Autonomous Greenhouse Challenge logged climate traces; and scores the same components on logged greenhouse data.

在 GreenLight-Gym 中，该框架将标量奖励分解为温度、CO2、湿度与饱和水汽压差（VPD）、遮阳幕布以及执行器代理项等条件项；将 GreenLight 适配到第二届自动温室挑战赛记录的气候轨迹中，并对记录的温室数据中的相同组件进行评分。