Landscaping Agricultural and Animal Husbandry Production Park Using Lightweight Deep Reinforcement Learning under Circular Symbiosis Concept

Abstract

The paper intends to optimize the landscape of the agricultural and animal husbandry (AG and AH) production park using the deep reinforcement learning (DRL) model under circular symbiosis. Therefore, after reviewing the relevant literature, decision tree evolutionary algorithm, and ensemble learning criteria, this paper studies and constructs the circular symbiotic industrial chain. Then, an experiment of landscaping the park and optimizing the production is made with full consideration of practical institutions. Finally, the numerical results show that the yield of several crops has been significantly improved after the landscape optimization by the proposed DRL model. Remarkably, the increase in rice yield is the most prominent. The yield of rice and wheat was about 12 kg before optimization and 18 kg after DRL model optimization, which has increased by 6 kg. This research has important reference value for improving the output efficiency of AG and AH products.