A Study on Vegetable Commodity Selling Price and Sales Forecasting Based on ARIMA-Differential Evolutionary Algorithm

Chengcheng Liu; Guangyin Zhou; Caiying Zhong

doi:10.54097/1gh4h491

Authors

Chengcheng Liu
Guangyin Zhou
Caiying Zhong

DOI:

https://doi.org/10.54097/1gh4h491

Keywords:

Vegetable Commodity, Selling Price, Sales Forecast, ARIMA, Differential Evolutionary Algorithms.

Abstract

Fresh food superstores need to develop accurate replenishment plans to cope with the pricing challenges caused by the short freshness period, many varieties, and large differences in origins of vegetables, to meet customer demand and to protect the sales program. In this paper, we firstly draw bar charts to show the distribution of sales among the top 10 best-selling and least-selling vegetable items in each vegetable category, and draw correlation coefficients heat maps to show the correlation of sales among the six vegetable categories. Since there are too many vegetable categories, only the 10 best-selling individual items are selected to draw the heat map of correlation coefficient, and hierarchical clustering is performed to study the correlation. Then a cost-plus pricing model is established to represent the recommended selling price through the total cost per unit and profit margin, while the total sales volume of each category is obtained by accumulating the sales volume of each individual item. In this paper, K-means clustering was used to aggregate the six vegetable categories into three broad categories, and it was found that the sales volume and the recommended selling price were negatively correlated. In order to predict the daily replenishment volume of each category from July 1-7, 2023, this paper screens the daily sales volume of the categories in June and July in the past two years, and predicts it with the ARIMA model in time series analysis, and draws line graphs to show the prediction results, and gives the daily replenishment volume strategy at the same time. Finally, differential evolutionary algorithm is used to introduce the optimal selling price, which is 39 yuan/kg for aquatic roots and tubers, 24 yuan/kg for flowers and leaves, 17 yuan/kg for cauliflower, 3 yuan/kg for eggplant, 26 yuan/kg for chili peppers, and 37 yuan/kg for edible mushrooms. The maximum gain was 73,525.45 yuan.

Downloads

Download data is not yet available.

References

YANG Yue, FENG Gu Cheng, LI Xin Ye, et al. Internet+" sales model of celery products [J]. Changjiang Vegetable, 2020 (20): 4.

LI Zongyou, ZHAO Xiaoli, MA Yinyi, et al. Research on automatic pricing and replenishment decision problem for vegetable commodities [J]. Advances in Applied Mathematics, 2023, 12 (11): 14.

A Qi. Research on Pricing of Scientific and Technological Research Services Based on Cost-plus Pricing Method [J]. Library Research and Work, 2021 (10): 8.

Lv Wang. Research on trend prediction of fresh vegetables sales based on improved SVM [J]. [2024-03-11].

Zhang Yanliang, Liu Ya. Research on fresh e-commerce product sales prediction based on fractional order gray model [J]. Preservation and Processing, 2021, 21 (10): 6.

WU Jiuniu, GAO Decheng, JIANG Weidong, et al. Application of interpolation method based on box-and-line diagram in data processing of empty box barometer [J]. Industrial Instrumentation and Automation Device, 2023 (3): 93 - 98.

ZHANG Huifeng, JING Tian. The combination of cost plus and consumer perceived value: A pricing model for knowledge payment products [J]. Journalism and Communication Research, 2021 (1): 38 - 51, 127.

Zhu J, Wang H. An Improved K-means Clustering Algorithm [J]. Microcomputer Information, 2010, 10 (1): 193 - 199.

Tao Z F, Tan W F, Chen H Y. A time series prediction method for component data incorporating fuzzy time series analysis [J]. Systems Engineering Theory and Practice, 2023, 43 (5): 1534 - 1544.

Ge Yuanyuan, Chen Depo, Zou Feng. Multi-cluster multi-strategy differential large-scale multi-objective optimization algorithm [J]. Control and Decision Making, 2024 (002): 039.