Artificial neural network as a support of platform track assignment within simulation models reflecting passenger railway stations

Abstract

Simulation models reflecting the operation of transportation systems are supposed to utilize an appropriate realization of decision-making procedures. The formation or selection of adequate modelling approaches and methodologies has to be done in this context. Attention is paid to a specific operational problem related to the assignment of platform tracks to delayed arriving trains within simulation models of passenger railway stations. The application of a twolayered artificial neural network as a decision-making support associated with the mentioned assignment problem was investigated. The neural network reached very encouraging results with regard to the studied problem, which enables its profitable utilization. It means, in fact, that the quality of relevant simulation models rises and the consequent credibility of corresponding simulation studies for railway companies should be increased.Simulation models of passenger railway stations include quite a wide range of decision-making tasks, which deal mainly with the following.JRRT268

Research on Operation Arrangement for Waiting Hall in Railway Passenger Station

“…The end of time slices (min) block[0] -17.5 20 block [1] 24.6 31 block [2] 83.5 96 block [3] 282.1 314 block [4] 342.3 351 block [5] 358.9 367 block [6] 413.5 439 block [7] 453.9 463 block [8] 503.4 516 block [ 9] 519.6 548 block [10] 550.3 577 block [11] 578.6 625 block [12] 636.7 648 block [13] 693.9 781 block [14] 805.7 824 block [15] 826.2 855 block [16] 862.5 885 block [17] 887.6 910 block [18] 914.6 921 block [19] 923.1 1006 block [20] 1009.4 1088 block [21] 1089 1104 block [22] 1110 Operation arrangement of parking lines in railway large-scale passenger technology stations is a complicated system engineering that is influenced by multitudinous factors, such as the scheduling of shunting locomotive and mutual interference in throat area and so on, which has a direct influence on the efficiency of railway passenger technology stations.…”

Optimal operation scheduling and platform track assignment in a passenger railway station

“…An approach based on the use of an artificial neural network to solve the problem of platform track assignment for a delayed train at its arrival at the station was introduced in Bazant 10 and Bazant and Kavicka. 11 Another approach is to use the decision-making support tool ROMA (Railway traffic Optimisation by Means of Alternative graphs). 12 As soon as regular operation is disturbed, new train routes (re-routing) are computed using the tabu search method.…”

Keleivių vežimo geležinkeliu priemonių parinkimo tyrimas

“…Dėmesys skiriamas specifinei eksploatacinei problemai, susijusiai su platformos kelių paskyrimu vėluojantiems traukiniams keleivinių geležinkelio stočių imitaciniame modelyje. Tirtas (Bažant, Kavička 2009) dvisluoksnės Neural Network technologijos naudojimas darant sprendimus minėtai kelio paskyrimo problemai spręsti. Nagrinėjant nurodytą problemą ši Neural Network programa parodė daug žadančius rezultatus, dėl kurių galimas pelningas jos naudojimas.…”