The railway infrastructure plays a pivotal role in fostering economic growth and poverty alleviation. Over time, these infrastructures deteriorate due to aging and usage, compromising their functionality. Given their crucial socioeconomic significance and vast scale, ensuring their functionality and availability is paramount. Thus, an effective condition-based maintenance program is essential to restore reliability, facilitate cost-effective restoration, and enable continued benefits.

Track is a critical railway component susceptible to degradation from traffic loading, resulting in deviations from designated geometry parameters. Such degradation jeopardizes safety, availability, and travel quality. Developing an effective tamping regime emerges as a vital maintenance measure to control degradation and restore track geometry to acceptable standards. An optimal maintenance schedule becomes imperative to minimize costs, enhance track availability and capacity, and ensure safety.

Achieving efficient tamping scheduling necessitates accurate prediction of geometry degradation, accounting for tamping effects, and modeling the evolution of single defects. However, uncontrolled shifts in geometry measurements from different inspections—known as positional errors—can misplace defects and distort their evolution analysis. Therefore, precise alignment of geometry measurements is vital to eliminate such positional errors.

The purpose of this research was to streamline maintenance scheduling through leveraging track geometry measurements for modeling and prediction. Firstly, a study addressed alignment through evaluating and comparing four methods—Cross-Correlation Function (CCF), Recursive alignment by fast Fourier transform (RAFFT), Correlation optimized warping (COW), and Dynamic Time Warping (DTW). Furthermore, a combined RAFFT-COW method was proposed, overcoming their limitations. Comparison revealed COW aligned datasets satisfactorily without altering their shape but could not align endpoints precisely. The combined method effectively aligned datasets even when the datasets were stretched or compressed. Secondly, a modified COW (MCOW) addressed accurate and efficient alignment. MCOW surpassed COW's restrictions and reduced alignment time. To enhance robustness, MCOW with channel fusion (MFCOW) combined data from different channels, significantly reducing positional errors. Thirdly, a multi-objective approach proposed aimed at reducing positional errors in geometry measurements of track as a linear asset. Accordingly, recursive segment-wise peak alignment (RSPA) and MCOW were evaluated and compared. Furthermore, a novel rule-based approach proposed which prevented data loss during alignment, preserving all the single defects. In addition, the results revealed that RSPA excelled in aligning peaks, while MCOW proved efficient for datasets with equal priority data points.

Finally, an optimization model minimized track geometry maintenance costs through tamping scheduling. Key track quality indicators, including the standard deviation of the longitudinal level and single defects and the impact of preventive/corrective tamping on these indicators were integrated. Results showcased the influence of fixed maintenance window costs and maintenance cycle intervals on tamping expenses. The model's validity was confirmed through interactions with experienced practitioners from prominent railway infrastructure and maintenance entities.