Aiming at the problem of low prediction precision of the existing Internet of Things（ IoT） environment temperature prediction methods and lag of the prediction results

an IoT environment temperature prediction model based on optimized proximal policy optimization（ PPO） algorithm and Actor-Critic（ AC） network is proposed

PPOAC model

which combines the AC reinforcement learning network to construct a dual network model for temperature prediction

and uses the optimized PPO algorithm to dynamically select the loss function.Finally

the IoT ambient temperature dataset provided by the Kaggle data platform is used to verify the effectiveness and stability of the model through experiments.