# Optimization decision model of vegetable stock and pricing based on TCN-Attention and genetic algorithm

**Authors:** Linhan Xia, Jinyuan Zhang, Bohan Wen  
**arXiv:** [2403.01367](https://arxiv.org/abs/2403.01367)  
**Date:** 2024-03-03  
**Summary page:** https://www.summarizepaper.com/en/paper/2403.01367v1/

## Key points

- Study focuses on optimizing decision-making process for vegetable stock and pricing in supermarkets
- Importance of informed decisions regarding procurement costs and allocation quantities of vegetables for profitability
- Analysis of relationship between pricing and allocation strategies faced by supermarkets in vegetable operations
- Use of advanced techniques such as optimization algorithms, linear regression, TCN-Attention model, TOPSIS evaluation model, genetic algorithm
- Integration of historical data on vegetable costs with time information based on 24 solar terms to train cost prediction model
- Identification of 32 most market-demanded products using TOPSIS evaluation model
- Utilization of genetic algorithm to search for globally optimized vegetable product allocation-pricing decisions
- Accurate daily cost prediction of single vegetable products using TCN-Attention model combined with 24 solar terms information
- Consideration of time series factors, seasonal factors, influence of dates on vegetable crops when predicting purchasing costs
- Use of entropy weight method to objectively reflect importance of each index in evaluation process while avoiding errors from subjective weighting

## Summary

The study focuses on optimizing the decision-making process for vegetable stock and pricing in supermarkets. As the vegetable market in China continues to grow rapidly, it has become crucial for supermarkets to make informed decisions regarding procurement costs and allocation quantities of vegetables to ensure profitability. The paper analyzes the relationship between pricing and allocation strategies faced by supermarkets in their vegetable operations. Advanced techniques such as optimization algorithms and linear regression are employed to determine replenishment and pricing strategies. Historical data of 61 different products is used to establish a clear relationship between sale prices and sales volumes. By integrating historical data on vegetable costs with time information based on the 24 solar terms, a cost prediction model is trained using TCN-Attention. The study also introduces the TOPSIS evaluation model which identifies the 32 most market-demanded products. A genetic algorithm is then utilized to search for globally optimized vegetable product allocation-pricing decisions. The TCN-Attention model combined with information on the 24 solar terms accurately predicts daily costs of single vegetable products. Time series factors are taken into account by introducing seasonal factors and considering the influence of dates on vegetable crops based on the 24 solar terms when predicting purchasing costs. Additionally, an entropy weight method is used to objectively reflect the importance of each index in the evaluation process while avoiding errors caused by subjective weighting. While there are some identified shortcomings in this study, they can be addressed and optimized in future research endeavors. Overall, this research provides valuable insights into optimizing supermarket operations in selling vegetable products through strategic decision-making processes supported by advanced algorithms and models.