Graphical neural network-based MCDM for evaluating sustainable city logistics measures

This paper presents a novel approach to evaluate sustainable city logistics measures using a multi-criteria decision-making (MCDM) framework based on qualitative data. The proposed approach combines different weighing methods such as fuzzy DEMATEL, fuzzy AHP, to determine the weights of the criteria. The criteria are based on economic, environmental, social, and technical aspects, such as operational costs, energy consumption, revenues, air pollution, noise, land use, congestion, accidents, mobility, accessibility, freeing of public space, logistical efficiency, trip effectiveness, loading factor of vehicles, service quality, and customer coverage'. The study also proposes a graphical neural network (GNN) with weighted MCDM for artificial intelligence-based decision-making, which helps in achieving a better selection of city logistics measures for urban freight logistics. The approach is implemented in MATLAB and compared with other deep learning models such as CNN, ANN, and DNN, showing higher performance. Furthermore, the paper presents an extrapolation-enhanced approach for modelling travel decision-making based on MCDM with GNN and error calculation, which provides higher performance. The results demonstrate the effectiveness of the proposed approach in selecting sustainable city logistics measures, which can contribute to improving urban freight logistics.