Data-Driven Machine Learning System for Optimization of Processes Supporting the Distribution of Goods and Services – a case study

Abstract

The article presents the concept of a data-driven machine learning system increasing the efficiency of the process of supporting the distribution of goods and services, which arises as a result of the R&D project implemented within the framework of the Operational Programme Smart Development (EU) 2014-2020, financed by National Centre for Research and Development in Poland. The solutions obtained within the presented concept will eventually become components of the Distribution Optimization System, which will be applied on the transport and logistics services market. The elements of the presented concept are geocoding algorithms (i.e. precise calculation of the delivery address, based on the transmitted information), route calculation algorithms (solving the modified VRPTW problem in stochastic time-dependent networks) and distribution region optimization algorithms, as well as smart tracking device monitoring vehicle traffic parameters, characterized by extremely long autonomous operation time (on batteries) and resistance to GPS interference.