Multi-Layer Data Model for Transportation Logistics Solutions

• Main Authors: Anton Ivaschenko, Sergey Maslennikov, Anastasia Stolbova, Oleg Golovnin
• Format: Article
• Language: English
• Published: FRUCT 2020-04-01
• Series: Proceedings of the XXth Conference of Open Innovations Association FRUCT
• Subjects: transportation logistics; data model; big data
• Online Access: https://www.fruct.org/publications/fruct26/files/Ivas.pdf
• ISSN: 2305-7254; 2343-0737

The paper presents an original multi-layer data model for software solutions to be used in transportation logistics. Based on an analysis of the conceptual model it is proposed to decompose the data into layers. As a basic decomposition approach, it is recommended to consider processing capacity characteristics and performance measures instead of objects and subjects classification typical for human perception. The proposed model allows increasing the efficiency of parallel computing algorithms implementation, specific for Big Data analysis. Two algorithms were proposed based on the method of criteria comparison. The sequential algorithm implements the classical approach to find a path on a graph, while the parallel one uses an approach that makes it possible to increase the efficiency of layer-by-layer task separation, taking into account the capabilities of a computing system for simultaneous parallel calculations. Experimental results prove the necessity to introduce original data structures for parallel processing of this data. The approach is implemented using Apache Spark and Stream API. A study conducted on real data on 21 settlements in the Netherlands showed an advantage in the execution time of the parallel computing algorithm over the usual sequential search of 38%.