Transport Service Scheduling in Remote Regions by the Application of Machine Learning

Transportation operations in sparsely populated or rural areas are planned and organized in an efficient manner. The objective is to maximize the efficiency of transport service scheduling in order to cater to the movement requirements of the inhabitants in these regions. The aim is to optimize the functionality and efficiency of transportation services in rural regions by coordinating the available vehicles in a manner that satisfies the needs and preferences of the local populace while also taking into account logistical limitations. The implementation of machine learning incorporates a theoretical and data-driven strategy to enhance the transit timetable decision-making process. A dataset including vehicle capacities, available vehicle counts, and desired travel hours for people is essential to the project code. This project use unsupervised machine learning methods since the data does not have any labels. Clustering is a popular unsupervised machine learning technique because it groups similar data into meaningful categories. Among the clustering techniques that were used were K-Medoids, DBSCAN, Mini-Batch K-Means, and K-Means. Hierarchical clustering is chosen because it can keep the total number of clusters and the number of nodes within each cluster at a constant value. Once the data points have been aggregated using the algorithms, the silhouette score is used to choose the approach that is most suited for vehicle scheduling. The relevance of cluster centers determines the final scheduling of the vehicles.