Application of Artificial Intelligence in Prediction of Road Freight Transportation

Other literature type, Article English OPEN
Mrowczynska, Bogna ; Ciesla, Maria ; Krol, Aleksander ; Sladkowski, Aleksander (2017)
  • Publisher: Faculty of Traffic and Transport Sciences, University of Zagreb
  • Journal: Promet - Traffic&Transportation, volume 29, issue 4 (issn: 0353-5320, eissn: 1848-4069)
  • Related identifiers: doi: 10.7307/ptt.v29i4.2227
  • Subject: forecasting methods; freight transport; artificial immune system; clonal selection; Bayesian networks; double exponential smoothing | TA1001-1280 | forecasting methods | freight transport | double exponential smoothing | clonal selection | artificial immune system | Bayesian networks | Transportation engineering

<p>Road freight transport often requires the prediction of volume. Such knowledge is necessary to capture trends in the industry and support decision making by large and small trucking companies. The aim of the presented work is to demonstrate that application of some artificial intelligence methods can improve the accuracy of the forecasts. The first method employed was double exponential smoothing. The modification of this method has been proposed. Not only the parameters but also the initial values were set in order to minimize the mean absolute percentage error (MAPE) using the artificial immune system. This change resulted in a marked improvement in the effects of minimization, and suggests that the variability of the initial value of S2 has an impact on this result. Then, the forecasting Bayesian networks method was applied. The Bayesian network approach is able to take into account not only the historical data concerning the volume of freight, but also the data related to the overall state of the national economy. This significantly improves the quality of forecasting. The application of this approach can also help in predicting the trend changes caused by overall state of economy, which is rather impossible when analysing only the historical data.</p>
  • References (13)
    13 references, page 1 of 2

    [1] Lawton R. How should additive Holt-Winters estimates be corrected? Int. J. Forecast. 1998;14(3):393-403. doi:10.1016/S0169-2070(98)00040-5

    [2] Pinto R, Gaiardelli P. Setting forecasting model parameters using unconstrained direct search methods: An empirical evaluation. Expert Syst. Appl. 2013;40(13):5331-5340. doi:10.1016/j. eswa.2013.03.044

    [3] Yager RR. Exponential smoothing with credibility weighted observations. Information Sciences. 2013;252:96-105. doi:10.1016/j.ins.2013.07.008

    [4] Lin K-P, Pai P-F, Yang S-L. Forecasting concentrations of air pollutants by logarithm support vector regression with immune algorithms. Appl. Math. Comput. 2011;217(12):5318-5327. doi:10.1016/j. amc.2010.11.055

    [5] Lénárt B. Automatic identification of ARIMA mod - els with neural network. Period. Polytech. Transp. 2011;39(1):39-42. doi:10.3311/pp.tr.2011-1.07

    [6] Eisuke K, Maasaki H, Takao M. Application of Bayesian Network to stock price prediction. Artif. Intell. Research. 2012;1(2):171-184.

    [7] Acar Y, Gardner ES. Forecasting method selection in a global supply chain. Int. J. Forecast. 2012;28(4):842- 848. doi:10.1016/j.ijforecast.2011.11.003

    [8] A guidebook for forecasting freight transportation demand. Cambridge Systematics, Inc. 1997; NCHRPreport 388, Transportation Research Board, National Research Council, Washington, D.C.

    [9] Wierzchoń ST. Sztuczne systemy immunologiczne. Teoria i zastosowania. Warszawa Akademicka Oficyna Wydawnicza EXIT; 2001. Polish.

    [10] Bolstad WM. Introduction to Bayesian statistics. Wiley-Interscience; 2004.

  • Metrics
    No metrics available
Share - Bookmark