Title: Machine learning for optimisation of flow-rack AS/RS performances

Authors: Zakarya Amara; Latefa Ghomri; Ali Rimouche

Addresses: Manufacturing Engineering Laboratory of Tlemcen, Abou Bekr Balkaid University Tlemcen, 13000 Tlemcen, Algeria ' Manufacturing Engineering Laboratory of Tlemcen, Abou Bekr Balkaid University Tlemcen, 13000 Tlemcen, Algeria ' Dynamical Systems and Applications Laboratory, Abou Bekr Balkaid University Tlemcen, 13000 Tlemcen, Algeria

Abstract: In this paper, we are interested in flow-rack automated storage/ retrieval systems (AS/RS), which are compact AS/RS. For this configuration of AS/RS we propose a new storage method based on machine learning (ML), i.e., ML method that assigns to each incoming load a position in the rack, in such a way, that the retrieval time of this same load will be optimal. In other words, we tidy out the loads inside the rack, In order to facilitate access to each type of loads. Consequently, the total (average) retrieval time in the system is minimised. The choice of ML is mainly due to the fact that the output, which is the minimisation of the average retrieval time, cannot be expressed as a function of the input, which is the choice of the most appropriate cell, for the storage of each incoming load. We compared the proposed model results with other basic storage methods. The obtained results were very satisfactory.

Keywords: flow rack AS/RS; retrieval time prediction; supervised machine learning; regression; classification.

DOI: 10.1504/IJISE.2024.137955

International Journal of Industrial and Systems Engineering, 2024 Vol.46 No.3, pp.390 - 403

Received: 09 Apr 2022
Accepted: 29 Jun 2022
Published online: 12 Apr 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article