Research On Green Warehouse Optimization For Company C Based On Combination Demand Forecasting

The important tasks in warehouse management and its optimization within the supply chain are the demand forecasting for future product sales and the significant impact on carbon emissions from logistics activities and processes across all companies in the supply chain.With the increasing diversification of demand in the fast-moving consumer goods market,companies have gradually shifted their production and warehousing operations towards market-oriented approaches.Demand forecasting has also evolved from overall quantity predictions to category-based and individual product predictions,incorporating new influencing parameters.Single-factor demand forecasting methods and conventional green indicators are no longer sufficient to assist fast-moving consumer goods companies in accurately capturing the market and improving green warehouse management.Therefore,this paper takes a practical perspective and proposes a demand forecasting model for a combination of fast-moving consumer goods suppliers,along with an optimized approach for specific company’s green indicators.The study selects Company C as a case study,analyzing relevant data on cigarette sales and warehousing to validate the effectiveness of the proposed method.The specific content is as follows:1)The research focuses on a combination forecasting method based on the Seasonal Autoregressive Integrated Moving Average model-Sparrow Search Algorithm-Radial Basis Function Neural Network(referred to as SARIMA-SSA-RBF).Since the monthly sales data for products exhibit seasonality as a time series,the analysis of market demand forecasting needs to consider their seasonal characteristics.First,the Seasonal Autoregressive Integrated Moving Average model(SARIMA)is applied to forecast the market demand.Simultaneously,the seasonal index and residual sequence of the product’s sales data are extracted as data preparation for the next step of forecasting.The planned sales parameters and the residual sequence are then combined using the Sparrow Search Algorithm-Radial Basis Function Neural Network(SSA-RBF)to fit and forecast the residuals.Taking Company C’s Jiaozi(Blue)cigarette specification as an example,the two forecasting methods are compared,and the Mean Absolute Percentage Error(MAPE)for the combined forecasting method is calculated as10.31%,indicating a significant improvement in prediction accuracy compared to the SARIMA model.2)The study focuses on the simplification of green indicators and carbon emissions calculation.Referring to "Construction and Accounting Methods of Green Logistics Indicators" and Company C’s assessment standards,the rough set reduction method is utilized to formulate green warehouse management indicators that align with Company C’s development needs.Carbon emissions calculations for Company C are also performed,determining values under various indicators.By analyzing three in-sale products of Company C and their shared warehouse total sales volume,the combination model is verified to improve the accuracy of single-product forecasts and reduce carbon emissions generated during the logistics process.Through verification,the SARIMA-SSA-RBF combination forecasting model,when adopted by Company C,achieves a single-product forecast accuracy of up to 89.69%,representing an improvement compared to the previous forecasting accuracy.With improved accuracy in individual product forecasts,the company can align shared inventory and the allocation of storage locations for individual products more closely with market realities.This refinement of inventory management for each product enhances the turnover rate of shared warehouses.Additionally,it provides new ideas and directions for effective collaboration between suppliers and retailers.The methods proposed in this paper can serve as a reference for demand forecasting and the development of green warehousing indicators in various industries’ supply chain management.