ROMJIST Volume 23, No. T, 2020, pp. T41-T56
Emir ZUNIC, Dzenana DONKO Innovative Concept for Setting Up and Adjustment the Parameters of Real-world Vehicle Routing Problems: Case Study in Logistics
ABSTRACT: Transportation occupies one-third of logistics costs, and accordingly transportation systems largely influence the performance of the logistics system. Most of the algorithms for successful solving of vehicle routing problem (VRP) are consisted of several control parameters and constants, so this paper presents the data-driven prediction model for adjustment of the parameters based on historical data, especially for practical VRP problems with realistic constraints in the field of freight and logistics. The concept is consisted of four prediction models: (1) Generalized Linear Models (GLM); (2) Support Vector Machine (SVM); (3) Decision Tree (DT); (4) Naive Bayes (NB), and decision support systems for comparing acquired results for each of the used models. Additionally, one of the parameters necessary for successfully solving any of the VRP problems is time of unloading goods (service time) for each observed customer. A novel approach for setting up the customers’ time of unloading goods is established. The successful feasibility of the given transportation routes, in a real environment, is also presented.KEYWORDS: Data Mining and Analysis, Data-driven concept, Vehicle Routing Problem, Parameter Setting Problem, Real-world datasetRead full text (pdf)