APPLICATION OF ARTIFICIAL INTELLIGENCE IN SINGLE WAGON LOAD (SWL) OPTIMIZATION OPERATIONAL MODEL FOR RAILWAYS OF REPUBLIKA SRPSKA

Abstract

In conditions of stagnation of the European Single Wagon Load (SWL) system, which records a market share decline from 50% to 35–36% between 2004 and 2011 (European Commission, 2015), and with the dramatic example of DB Cargo recording annual losses of ~€442 million directly attributed to SWL operations (2022), this paper investigates the possibility of introducing an AI-optimized SWL service on the Railways of Republika Srpska (ŽRS) network. Unlike established European operators burdened by inherited costs, ŽRS has no SWL history — which is positioned in this paper as a strategic advantage, not a disadvantage. The proposed model is based on four AI components: (1) demand aggregation module, (2) real-time transport chain optimization following the SBB Cargo TK-Module model (Moldenhauer et al., 2025), (3) ETA prediction for end users, and (4) backhaul optimization algorithm.