In today's context, it is undeniable that water transport, as an infrastructure industry, should be developing at a rapid pace. According to experts, in 2020-2022, it is expected that shipments by sea and annual vessels will increase to 2500 million tons; cargo processing at state-owned commercial ports will also increase to around 240 million tonnes; passenger traffic will amount to more than 10.9 million passengers.
Research done within the framework of this dissertation shows that the level of accident-free navigation, the quality and efficiency of passenger and cargo transportation continue to improve. Particular attention is given to measures to ensure the required level of reliability of technical means. A significant role in this direction is the decision of the task of forecasting the technical condition of the means of water transport. This further underscores the relevance of scientific research to improve the operational efficiency of marine and annual vehicles by utilizing information on the forecast of their technical condition.
At present, progressive hardware and software solutions are actively used to ensure the guaranteed level of reliability of prediction of the technical state of water transport facilities. The analysis of foreign and domestic experience in the development and implementation of systems for predicting the technical condition of both component subsystems and means of water transport in general indicates the possibility of a significant increase in their efficiency due to the development of mathematical and algorithmic support. The most relevant in this area is the use of models and methods of artificial intelligence, namely, the so-called soft computing. Despite the rapid development of artificial intelligence theory in general, the challenge is to improve the existing and develop new mathematical and software computer systems for predicting the technical state of water transport based on soft calculations.
Thus, in solving the problems of development of the transport industry of Ukraine, water transport in particular, the scientific task of improving the existing and development of new models and methods of forecasting the technical condition of water transport facilities on the basis of soft calculations is urgent, and this dissertation is devoted to solving this problem.
Scientific novelty of the obtained results is that:
for the first time developed a conceptual model for predicting the technical state of water transport, based on OLAP technology with intelligent data analysis on soft calculations, namely, the complex use of fuzzy clustering based on methods of subtractive clustering (to determine the number of values of linguistic variables) and fuzzy variables medium (FCM, Fuzzy C-Means) for building accessory functions, as well as the adaptive neural fuzzy ANFIS network. The software implementation of the model significantly improves the efficiency and reliability of the forecast by reducing the number of operations and the use of new principles of processing a priori information and self-study;
the model of forming the function of membership in the prediction of the technical state is improved, which, unlike the existing ones, uses fuzzy clustering: the method of subtractive clustering to determine the number of clusters that are interpreted as the rank of the base term-set of linguistic variables and the method of fuzzy c-means for the calculation; this greatly improves the adequacy of the phasing procedure and accuracy;
improved method for predicting the technical state of the FTA based on fuzzy inference when using a hybrid network, which, unlike the existing ones, is based on the model of adaptive neural fuzzy network system ANFIS (Adaptive Neuro-Fuzzy Inference System), which provides the opportunity to simultaneously use the benefits of fuzzy logic to improve the accuracy of predicting the technical condition.
