Język polski English
LogForum Logo
Scopus Logo
Webofsc Logo

ISSN 1895-2038, e-ISSN:1734-459X

Submit manuscript
Newsletter subscription
Journal metrics
Indexed in:
Creative Commons licence CC BY-NC (Attribution-NonCommercial)

Issue 4/ 2020, article 5

Rafael Granillo-Macías



Background: In the current economic scenarios, characterized by high competitiveness and disruption in supply chains, the latent need to optimize costs and customer service has been promoted, placing inventories as a critical area with high potential to implement improvements in companies. Appropriate inventory management leads to positive effects on logistics performance indices. In economic terms, about 15% of logistics costs are attributed to warehousing operations. With a practical approach, using a case study in a company in the food sector, this article proposes an inventory classification method with qualitative and quantitative variables, using data mining techniques, categorizing the materials using variables such as picking frequency, consumption rates and qualitative characteristics regarding their handling in the warehouse. The proposed model also integrates the classification of materials with techniques for locating facilities, to support decision-making on inventory management and storage operations.

Methods: This article uses a method based on the Partitioning Around Medoids algorithm that includes, in an innovative way, the application of a strategy for the location of the optimal picking point based on the cluster classification considering the qualitative and quantitative factors that represent the most significant impact or priority for inventory management in the company.

Results: The results obtained with this model, improve the routes of distributed materials based on the identification of their characteristics such as the frequency of collection and handling of materials, allowing to reorganize and increase the storage capacity of the different SKUs, passing from a classification by families to a cluster classification. Furthermore, the results support decision-making on storage capacity, allowing the space required by the materials that make up the different clusters to be identified.

Conclusions: This article provides an approach to improving decision-making for inventory management, showing a proposal for a warehouse distribution design with data mining techniques, which use indicators and key attributes for operational performance for a case study in a company. The use of data mining techniques such as PAM clustering makes it possible to group the inventory into different clusters considering both qualitative and quantitative factors. The clustering proposal with PAM offers a more realistic approach to the problem of inventory management, where factors as diverse as time and capacities must be considered, to the types and handling that must be had with the materials inside the warehouse.

Keywords: cluster, Partitioning Around Medoids, facility location, supply chain

Full text available in in english in format: Adobe Acrobat pdf article nr 5 - pdf

Streszczenie w jezyku polskim Streszczenie w jezyku polskim.

DOI: 10.17270/J.LOG.2020.512
For citation:

MLA Granillo-Macías, Rafael. "Inventory management and logistics optimization: A data mining practical approach." Logforum 16.4 (2020): 5. DOI: 10.17270/J.LOG.2020.512
APA Rafael Granillo-Macías (2020). Inventory management and logistics optimization: A data mining practical approach. Logforum 16 (4), 5. DOI: 10.17270/J.LOG.2020.512
ISO 690 GRANILLO-MACíAS, Rafael. Inventory management and logistics optimization: A data mining practical approach. Logforum, 2020, 16.4: 5. DOI: 10.17270/J.LOG.2020.512
EndNote BibTeX RefMan