A Study on Farmland Ecosystems Based on the Logistic Growth Model and XGBoost Algorithm

Abstract

This study focuses on the issue of chemical pesticide dependency in the transformation of agricultural ecosystems, using the logistic growth model and XGBoost algorithm as core analytical tools to construct a multi-dimensional model revealing ecosystem dynamics and the feasibility of organic agriculture. The agricultural food web model constructed based on the logistic growth model uses bats and birds as top consumers, covering the complete food chain structure from producers to tertiary consumers. By expanding plant, insect, bird, and bat populations, as well as a soil nutrient model, to systematically analyse the impact of herbicides and insecticides on biological populations and the predation and competition relationships between species. The study found that the use of chemical pesticides leads to a decline in the populations of various biological species and soil nutrient content, triggering fluctuations in ecosystem stability. Following validation of ecological restoration potential through species regression models, the study employs the XGBoost algorithm to construct a multi-objective optimisation model, with economic profit and ecological sustainability as core objectives, combined with constraints such as budget and yield, to provide farmers with a framework for assessing the feasibility of organic agriculture. The study aims to form a complete research loop from ecological mechanism analysis to production practice decision-making, providing theoretical and methodological support for addressing imbalances in farmland ecosystems.