A hybrid model for forecasting the volume of passenger flows on Serbian railways

The accuracy of predicting the volume of railway passenger flows is very significant because of the vital role in the basic functions of transportation resources management. Although dealing with this problem is very often based on the use of the neural networks, the uncertainty which dominates in t...

Full description

Saved in:
Bibliographic Details
Published in:Operational research 2016-07, Vol.16 (2), p.271-285
Main Authors: Glisovic, Natasa, Milenkovic, Milos, Bojovic, Nebojsa, Svadlenka, Libor, Avramovic, Zoran
Format: Article
Language:English
Subjects:
Business and Management
Computational Intelligence
Forecasting
Genetic algorithms
Learning theory
Literature reviews
Management Science
Mathematical models
Neural networks
Operations Research
Operations Research/Decision Theory
Original Paper
Passengers
Planning
Railroad transportation
Railway networks
Railways
Studies
Support vector machines
Time series
Transportation
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The accuracy of predicting the volume of railway passenger flows is very significant because of the vital role in the basic functions of transportation resources management. Although dealing with this problem is very often based on the use of the neural networks, the uncertainty which dominates in the functioning of transportation systems is of great significance. The neural networks have been used for the time-series prediction with good results. This research compared two methods the parametric and the non-parametric approach. This study aims at presenting a hybrid model based on the integration of the genetic algorithm (GA) and the artificial neural networks (ANN) for forecasting the monthly volume of passengers on the Serbian railways. This innovative hybrid demonstrates how the genetic algorithms can be used to optimize the network architecture. By applying the idea of genetic algorithms in the neural networks, the integration is used so that on the basis of the input data, the selected population represents the number of neurons in the middle. In order to assess performances, the developed approach is compared to the traditional SARIMA model and the proposed method GAANN is better.
ISSN:1109-2858
1866-1505
DOI:10.1007/s12351-015-0198-5