Estimation of Route Reliability in Multimodal Hierarchical Hub Location Problem Using Lagrangian Relaxation and Artificial Neural Networks (ANN)

Abstract

The present study has developed an Artificial Neural Networks (ANN) model to predict the route reliability in various structures of the Multimodal Hierarchical Hub Location Problem (MHHLP) utilizing Lagrangian relaxation; so that, initially, a mixed integer programming model was proposed for MHHLP and the, an efficient Lagrangian relaxation method was developed to solve the problem in different structures. The results obtained from problem solving were used as input and output data to create an ANN model using Multilayer Perceptron (MLP) neural network. As a result, an ANN model was designed by which, the reliability of the MHHLP route was predicted in large dimensions at different values ​​of the parameters. Computational analysis, ANN model validation and prediction process were conducted using CAB and IAD data.