Abstract
An improved BP neural network algorithm for food demand prediction based on supply chain management is presented to realize market and sale management target effectively for food enterprises. First, the working principle of BP neural network algorithm is analyzed to explore the root reasons of its low convergence speed; Second, the paper integrates genetic algorithm with BP algorithm to present a new algorithm, then improves it through encoding chromosome, formatting fitness function, designing selection operator, redesigning crossover operator, designing mutation operator, integrating BP algorithm and optimal individual, improving calculation process step by step; Finally, a supply chain of a food enterprise is taken for experimental sample to illustrate the calculation performance of the improved algorithm and the simulation results indicate that the improved algorithm not only can solve the problem of low convergence speed, but also can improve the demand prediction accuracy and can be used for predicting supply chain demand for food enterprises practically.
Highlights
In the food enterprise, inventory is a reservoir to regulate the imbalance between the production and demand
With the increasing competition in current market, the rapid high-tech development, shrink of food product life cycles and the complexity of food product mix, coupled with the various characteristics and storage requirements of different food, all of these factors, affect supply chain and require increasing demand prediction accuracy to avoid losses caused by shortages of stock, which increases the cost of supply chain
Genetic algorithm is adopted in the paper to confirm the initial parameter value of BP algorithm to overcome the defect of BP network, that is falling into local optimization
Summary
Inventory is a reservoir to regulate the imbalance between the production and demand. The popular demand prediction methods for food supply chain adopted by enterprises and researchers are can divided to two categories: one is single prediction method, such as grey prediction method, adopting neural network prediction method, markov prediction method, prediction method based on quantity of value, time series prediction method; Another methods is using multiple prediction methods and combing their results according to certain form, i.e., combination prediction method. There are researchers presents some other methods, such as scenario analysis method Specific methods have their own advantages and disadvantages: Delphi method: Is always used for new product and long-term prediction, technology prediction and profits prediction. Its advantages includes it can draw on the wisdom of the masses and be beneficial to a reliable and comprehensive prediction. Epitaxial smoothing technique, when the impact of Fig. 1: The working topology of BP neural network algorithm seasonal fluctuation and change trend on prediction value are under considered and the seasonal
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
