Predicción de la producción y optimización de recursos en fideos victoria con Machine Learning y Algoritmos de optimización

Abstract

In the food sector, the need to modernize processes using digital technologies has become a priority in order to improve competitiveness. Faced with this challenge, this project proposes the development of an intelligent system, based on machine learning, to predict production at the company Fideos Victoria. This solution responds to the urgent need to move away from empirical practices, proposing an automated and adaptive approach.

The system designed is capable of analyzing the company's historical data and sales patterns using machine learning, which allows it to anticipate future demand and plan daily and weekly production based on this. In addition, it integrates a distribution route optimization module with the ACO algorithm, which reduces product delivery times and adapts logistics to market realities.

Through the implementation of this tool, the operational uncertainty of the production system was reduced by 84.06%, improving product availability. A significant improvement in predictive accuracy was also observed, with an average mean square error of 12.15% due to algorithm optimization for Random Forest. On the other hand, by applying the ACO algorithm, route variability was reduced by up to 1.57%, thereby improving the stability of the logistics system.

This technological advance represents an important step in the company's digital transformation and establishes a solid foundation for future similar applications in the food sector.