Research on Vegetable Cost Pricing and Sales Forecast Based on PSO Optimization and Forecast Model

Boyu Wu,Tian Wang,Ziting Zhu

DOI: https://doi.org/10.25236/ajbm.2024.060519

Abstract:: Due to the short shelf life of vegetable items in fresh produce superstores, most varieties need to be restocked on a daily basis. In order to make reasonable pricing and replenishment decisions, this paper firstly integrates data and establishes a multiple linear regression model to find out the relationship between price and sales volume, and further optimizes it using PSO particle swarm algorithm to obtain the linear relationship equation, and then establishes an ARIMA time-series prediction model using SPSS to predict the total amount of replenishment and the pricing strategy for July 1-7 by using the data of the last 4 weeks. When considering the number of sales category constraints, this paper carries out the single-product profit ranking and selects the top 33 single-products with higher profits by combining the available varieties on the basis of meeting the minimum display quantity, and predicts the replenishment and pricing strategy on July 1 using the GWO gray prediction model. Ultimately, based on the above conclusions, the strategy for superstores to maximize returns in pricing and replenishment is derived.

Business,Agricultural and Food Sciences

What problem does this paper attempt to address?

This paper attempts to address the issue of how to formulate reasonable pricing and replenishment strategies to maximize supermarket revenue given the short shelf life of vegetables. Specifically, the paper achieves this goal through the following steps:

1. **Data Integration and Model Building**: First, integrate the data and establish a multiple linear regression model to identify the relationship between price and sales volume.
2. **Model Optimization**: Use the Particle Swarm Optimization (PSO) algorithm to further optimize the multiple linear regression model and obtain the linear relationship equation.
3. **Time Series Forecasting**: Utilize SPSS software to establish an ARIMA time series forecasting model, predicting the replenishment volume and pricing strategy from July 1 to July 7 based on data from the past 4 weeks.
4. **Single Product Profit Ranking**: Considering the limitation on the number of sales categories, rank the single products by profit and select the top 33 products with higher profits based on available varieties.
5. **Grey Prediction Model**: Based on the Grey Prediction Model (GWO), predict the replenishment volume and pricing strategy for July 1.

Ultimately, the paper proposes that, under the premise of maximizing revenue, the pricing on July 1 from high to low should be wild lotus root powder, colored sweet pepper (2), red pepper, etc., and the replenishment volume from high to low should be Yunnan oil wheat vegetable, Yunnan oil wheat vegetable (portion), broccoli, etc. In summary, this paper provides a reference for relevant pricing strategies for supermarkets by establishing the Particle Swarm Optimization algorithm and ARIMA time series model, and it has certain application value.