New Spatial Analysis and Hybrid Heuristics Enhance Truck Freight Tonnage Estimation Based on Weigh-in-Motion Data

Abstract

This paper presents a novel and practical methodology for freight tonnage estimation by leveraging two complementary datasets: Telemetric Traffic Monitoring Sites (TTMS) data and Weigh-In-Motion (WIM) systems. To estimate freight tonnage statewide and potentially nationwide with limited truck weigh-in-motion stations, we have proposed a multi-objective location-allocation model that associated TTMSs with WIM stations based on similar attributes. Additionally, we have developed a fuzzy k-prototype clustering-based non-dominated sorting genetic algorithm - simulated annealing algorithm (FKC-NSGASA) to solve the multi-objective location-allocation problem, enabling accurate estimation of truck volumes. To address the over-counting problem, we introduced a truck volume elimination method. Finally, we have aggregated annual truck tonnage using the truck volume data and the average tonnage of WIM stations. The proposed methodologies are validated using WIM data from 2012 and 2017 in Florida. The results demonstrate that our approach achieves higher estimation accuracy, showcasing its potential for accurately estimating statewide freight tonnage. Furthermore, the developed estimation framework and algorithm offer an effective and computationally efficient method for statewide freight traffic evaluation.