With the impact of global climate change and agricultural development, forecasting and analyzing cotton yields is crucial for agricultural planning and resource allocation. In order to provide a more accurate prediction of national cotton yield, a multi-objective locust optimal combination forecasting is proposed. Three single models, ARIMA time series model, Least Squares Support Vector Machine LSSVM model, and Recurrent Neural Network RNN model, are firstly applied to forecast the national cotton production data from 2009 to 2023. Then, a set of optimal solutions were obtained through the multi-objective locust iterative optimization process, and the single model prediction results were compared with the prediction results of the combined prediction method. It is verified through examples that the combined prediction method using multi-objective locust optimisation predicts results with smaller error and higher fitting degree, which proves that the model has good value in practical application and better reflects the actual changes of cotton production. Finally, using this method to forecast the national cotton production in 2024-2026 can provide a reference for the development of the cotton industry.

The protective effect is one of the key indicators of whether the mask can block the bacteria, and it is also the project that the mask is most prone to quality problems. This paper analyzes the influence of mask structure, ear band, filtration efficiency and multi-layer wearing on protective effect, providing some reference for optimizing mask structure design and improving mask quality.