Data-driven approach in predicting truck arrival time in logistics: a field study of India

Abstract

This study investigates the application of the Particle Swarm Optimization-tuned Gradient Boosting (GB-PSO) model for predicting truck arrival times. The proposed model incorporates optimized hyperparameters to enhance predictive performance, as measured by the coefficient of determination (R) and Root Mean Square Error (RMSE). Using a real-world logistics dataset, GB-PSO outperforms conventional Gradient Boosting (GB) and expected travel time estimations, achieving a higher R value and lower RMSE across training and testing datasets. The analysis of SHAP values highlights the dominant influence of transportation distance on model predictions. These findings validate the effectiveness of GB-PSO in practical logistics optimization, reducing error and improving reliability in time-sensitive transportation systems.