Please use this identifier to cite or link to this item:
Title: Gestión del inventario en una empresa del sector farmacéutico mediante algoritmos de Machine Learning
Author: de Mas Jaumot, Jordi
Director: Polo Navarro, Lorena
Tutor: Conesa Caralt, Jordi  
Keywords: Inventory level prediction
Machine Learning
Neural networks
Deep learning
Artificial Intelligence
Issue Date: 10-Jan-2021
Publisher: Universitat Oberta de Catalunya (UOC)
Abstract: Master¿s dissertation is about advanced inventory management so that the optimal stock level can be predicted based on machine learning algorithms. Prediction must be able to ensure that the level of stock will allow a good customer service, avoiding a stockout situation and the consequent loss of sales and reducing as far as possible the costs derived from inventory management. Data taken as the basis for this project is from the pharmaceutical sector. The prediction of the optimal stock is carried out in a single step, avoiding the need to know in advance the demand forecast and the distribution of its probability. The optimal inventory level is calculated directly by analysing the available data: sales history, real stock history, promotion data, holiday calendar, etc. Although in the pharmaceutical industry, specifically in manufacturers, there are several types of stock (raw materials, WIP, semi-finished products and finished products), the data provided for the project allows to focus on the finished products. Additionally, some factors with influence in the result of the prediction are considered, such as: seasonality or weekly patterns that show a higher sales level on certain days of the week. The development is carried out based on machine learning algorithms, analysing different types of them to decide the one that provides better performance in order to obtain an alternative method of stock management and ensure better customer service with the minimum cost for the company.
Language: Spanish
Appears in Collections:Bachelor thesis, research projects, etc.

Files in This Item:
File Description SizeFormat 
jde_masTFM0121.pdf2.38 MBAdobe PDFView/Open
jde_masPresentación0121.pdf1.09 MBAdobe PDFView/Open

This item is licensed under a Creative Commons License Creative Commons