Neural network model for predicting the execution time of a transport task
DOI:
https://doi.org/10.18372/2310-5461.49.15289Keywords:
transport system, transport task, freight forwarding company, intelligent technologies, neural networksAbstract
The article discusses the issue of obtaining a neural network model designed to predict the execution time of a transport task. The initial information for studying the model is the carrier’s data on the expected time for completing the task and the date of the trip. Using the Monte Carlo method, outgoing data was obtained with the help of which the neural network was trained. In the course of the study, the obtained results were analyzed, which indicate that the use of the developed neural network model for predicting the time of the transport task will significantly reduce the error in comparison with the estimate of the initial parameters. The basic principle of the logistic activation function is a value argument from a range, which in turn, can take any value. The article examines the scheme of information flows when performing one cycle of training neural networks using the Levenberg-Marquart algorithm. The developed methodology is intended to be the basis for the methodology for assessing the influence of the human factor on the time of the transport task. When predicting the execution time of a transport task under certain conditions, the largest of the simple approaches to this type of task, we used travel time calculation. In our article, we used the time obtained based on the length of the route and the average speed of the vehicle, as well as the average amount of time it takes for the forwarder to complete the tasks. The proposed solution to the transport problem does not always provide the required forecast accuracy, but we set ourselves the goal of creating a more modern neural network forecasting model, which in turn takes into account the calculations related to the criterion of seasonality and days of the week in the period when the task is supposed to be completed. The article is devoted to optimizing the structure of a neural network by using three datasets with rather large data arrays. At the same time in the absence of the necessary data array from the enterprise database, this array can be supplemented with the results of mathematical modeling at the preliminary stage of the study. The obtained results of the study allow using a neural network model to predict the execution time of a transport task, significantly reducing the forecasting error due to the route identifierReferences
Zaynyllina E.S., & Каynov А. S. (2007) Мeetodika podbora koeficientov pry reshenyi optimizatsyonnoy zadachy s pomotcuy neyronnoy sety Hopfiilda. Маteryaly vserossyiskoy naychnoy konferencyi stydentov and aspirantov «Моlоdyi issledovately – regyionam», Fevral 3-6, 2007, Т.1, Vologda, 103-105.
Меdvedev V. S., & Potyomkin V. G. Neyronnyi sety: ucheb. posob./ V. S. Меdvedev, V. G. Potyomkin; Matlab 6. Моskwa: Izd-vо DYALOG-МYFY, 2002. - 496 с.
Osigwe E., Yi-Guang L., Sampath S., Gbanaibolou J., Dieni I. (2017). Integrated Gas Turbine System Diagnostics: Components and Sensor Faults Quantification using Artificial Neural Network. 23rd International Society of Air Breathing Engines (ISABE) Conference – ISABE 2017, March 1-3, 2017, Manchester, 32-43.
Loboda I. O. (2010). Gas Turbine Condition Monitoring and Diagnostics. Gas Turbines. Science and Transport Progress. 119–144. doi: http://doi.org/10.5772/10210
Loboda, I. O., Feldshteyn, Y. E, Ponomaryov, V. A (2012). Neural Networks for Gas Turbine Fault Identification: Multilayer Perceptron or Radial Basis Network. International Jet Turbo Jet-Engines, vol. 29, Issue 1. 37–48. doi: http://doi.org/10.1515/tjj-2012-0005
Nakatsuji, T., Shibuya, S. (2019). Neural Network Models to Traffic Flow Problems. Neural Networks in Transport Applications. Science and Transport Progress, 249-262. doi: http://doi.org/10. 43249780429445286-12
Hussein, D., Rose, G. (2019). The Impact of Data Quantity on The Performance of Neural Network Freeway Incident Detection Models. Neural Networks in Transport Applications, 311-340. doi: http://doi.org/10. 4324/9780429445286-15
Kim, J., Bukhari, W., Lee, M. (2018). Feature Analysis of Unsupervised Learning for Multi-task. Convolutional Neural Network Journal, 783-797. doi: http://doi.org/10. 1007/s11063-017-9724-1
Profillidis, V.A., Botzoris, G.N. (2020). Artificial Intelligence-Neural Network Methods. Modelling of Transport Demand, 353-382. doi: http://doi.org/10.1016/b978-0-12-811513-8.00008-x
Pachomova, V.M., Konnov, M.S. (2020). Research of Two Approaches to Detect Network Attacks Using Neural Network Technologies. Science and Transport Progress, 81-93. doi: http://doi.org/10.15802/stp2020/208233
Zhao, T., Wang, J., Zhang, J. (2017). Real-time multimodal transport path planning based on a pulse neural network model. International Journal Simulation and Process Modelling, 351-356. doi: http://doi.org/10.1504/ijspm.2017.085559
Liu, J. (2013). Neural Network Sliding Mode Control. Radial Basis Function Neural Network Control for Mechanical Systems, 113-132. doi: http://doi.org/10.1007/978-3-642-34816-7_5
Liu, Y.S., Holmes, P., Cohen, J.D. (2008). A Neural Network Model of the Eriksen Task: Reduction, Analysis, and Data Fitting. Neural Computation, 345-373. doi: http://doi.org/10.1162/neco.2007.08-06-312
Romanenko, A.N. (2018). Feature combination for the task of neural network acoustic model learning. Scientific and Technical Journal of Information Technologies, Mechanics and Optics, 350-352. doi: http://doi.org/10.17586/2226-1494-2018-18-2-350-352
