Dynamic items delivery network: prediction and clustering

Mokhammad R Yudhanegara; Sapto W Indratno; Rr Kurnia N Sari

doi:10.1016/j.heliyon.2021.e06934

Dynamic items delivery network: prediction and clustering

Heliyon. 2021 May 8;7(5):e06934. doi: 10.1016/j.heliyon.2021.e06934. eCollection 2021 May.

Authors

Mokhammad R Yudhanegara¹, Sapto W Indratno¹, Rr Kurnia N Sari¹

Affiliation

¹ Statistics Research Division, Faculty of Mathematics and Natural Sciences, Institut Teknologi Bandung, Jalan Ganesha 10, Bandung 40132, Indonesia.

Abstract

Items delivery companies generally use a model to minimize delivery costs. From a mathematical perspective, the model is an objective function that involves constraints. Meanwhile, from a practical point of view, these constraints include aspects that affect item delivery, for example, delivery zones, number of delivery vehicles, vehicle capacity, trip routes, etc. However, the models built so far have not paid attention to changes in road density. This aspect can result in a nonoptimal delivery model, which results in not a minimum delivery cost. For this reason, this paper discusses how to divide zones using the clustering method and predict changes in the shipping zone of a dynamic network using predictive distribution. So, the model can work optimally if the delivery zones and delivery strategies are suitable.

Keywords: Dynamic network; Mathematics; Predictive distribution; Spectral.