Multi-criteria clustering analysis for large-scale public transport performance diagnosis

Abstract

Public transport is a key factor for the global economy; therefore, it has always been a directive of governments to report on its performance to authorities and public. The purpose of the present study is providing a large-scale performance diagnosis dashboard for bus public transport systems to deal with multi-criteria context. The proposed dashboard can assist transportation authorities in undertaking a comprehensive performance evaluation both at route and system level. The methodology of this study is an integration of (i) ordered multi-criteria clustering method based on the K-means algorithm and the FLOWSORT outranking method, (ii) weighted average and (iii) PROMETHEE parameters-based single-criteria analysis. Inspired by an interesting route level evaluation methodology from recent research, a template is generated to illustrate the proposed approach. Outcomes are promising for investing in other multic riteria clustering methods to deal with large-scale performance evaluation at both route and system levels. The proposed approach can fit any evaluation model based on performance criteria. It allows a detailed presentation of the diagnosis in spite of the large-scale context, which eases the optimization process.