|
2020. No. 2 Vol.14
|
Modeling of social and economic systems
|
7–20
|
This article deals with the application of transfer learning methods and domain adaptation in a recurrent neural network based on the long short-term memory architecture (LSTM) to improve the efficiency of management decisions and state economic policy. Review of existing approaches in this area allows us to draw a conclusion about the need to solve a number of practical issues of improving the quality of predictive analytics for preparing forecasts of the development of socio-economic systems. In particular, in the context of applying machine learning algorithms, one of the problems is the limited number of marked data. The authors have implemented training of the original recurrent neural network on synthetic data obtained as a result of simulation, followed by transfer training and domain adaptation. To achieve this goal, a simulation model was developed by combining notations of system dynamics with agent-based modeling in the AnyLogic system, which allows us to investigate the influence of a combination of factors on the key parameters of the efficiency of the socio-economic system. The original LSTM training was realized with the help of TensorFlow, an open source software library for machine learning. The suggested approach makes it possible to expand the possibilities of complex application of simulation methods for building a neural network in order to justify the parameters of the development of the socio-economic system and allows us to get information about its future state. |
|
21`–35
|
This paper offers one approach for solving a problem that arises for railway transport operators. The task is to manage the fleet of freight railcars optimally in terms of profit maximization. The source data for the transport operator is a list of requests received from customers, as well as the location of railcars at the beginning of the planning period. The request formed by each customer consists of departure station, destination station, name and volume of cargo that the customer would like to transport. The request also contains the rate that the customer has to pay to the transport operator for each loaded wagon transported. Planning is carried out for a month in advance and consists, on the one hand, in selecting the most profitable requests for execution, on the other hand – in building a sequence of cargo and empty runs that will fulfill the selected requests with the greatest efficiency. Direct transportation of loaded and empty railway cars is carried out by Russian Railways with pre-known tariffs and time standards for each of the routes. At the same time, tariffs for driving loaded wagons are additional costs for the customer of the route specified in the request (customers pay both the transport operator for the use of wagons and Russian Railways); transportation of empty wagons is paid by transport operators. To solve this problem, one of the possible ways to reduce it to a large-dimensional linear programming problem is proposed. An algorithm is proposed, the result of which is a problem written in the format of a linear programming problem. To demonstrate the approach clearly, a simplified problem statement is considered that takes into account only the main factors of the modeled process. The paper also shows an example of a numerical solution of the problem based on simple model data. |
|
36–47
|
The paper provides a brief description of cognitive management, which opens up unique opportunities for the effective management of enterprises in modern complex and unstable conditions. The problems of commercializing this promising paradigm are discussed. It is pointed out that the main, critical one of these problems is due to the lack of developed engineering of cognitive management. A conceptual framework for solving this problem is proposed, based on the convergence of the ideas and methods of the “cognitive school” and the empirical experience gained in knowledge engineering. The results of using the conceptual framework in four research projects of different industry orientations, with different internal conditions and different dynamics of the external environment are presented. The engineering prospects of the proposed framework are discussed in terms of the commercialization of the cognitive school identified by H. Mintzberg, B. Ahlstrand and D. Lampel 30 years ago. |
Data analysis and intelligence systems
|
48–63
|
The purpose of the research was to develop a market value appraisal methodology based on a set of a joint logarithmically normal distribution of price-forming factors. Joint logarithmically normal distribution means random vector component logarithms are distributed together jointly normally. This article suggests a method for appraising the real estate market value based on the statistical hypothesis of a joint logarithmically normal distribution and conditional distribution of prices with fixed values of pricing factors. The article suggests a method of offer price analysis from the point of view of its relevance to pricing factor values. We consider the features of the coefficient of development depending on the area of the land plot. Additional arguments are given in favor of estimating market value as a mode of conditional laws of price distribution. An example of a multidimensional log-normal distribution of prices and pricing factors such as the area of the improvements (improvements mean buildings and constructions) area and the land area in real data, i.e. for the case of a three-dimensional random vector. We present a formula for determining the absolute maximum density point of a multidimensional logarithmically normal random vector. The proof is given in the Appendix. The results obtained can be used to create information systems to support decision-making in valuation activities for real estate properties. |
|
64–83
|
One of the most dynamically changing parts of the labor market relates to information technologies. Skillsets demanded by employers in this sphere vary across different industries, organizations and even certain vacancies. The educational system in the most cases lags behind such changes, so that obsolete skillsets are being taught. This article proposes an algorithm of skillsets identification that allows us to extract skills that are needed by companies from different occupational groups in the information technologies sector. Using the unstructured online-vacancies database for the Russian regional labor market, skills are extracted and unified with the use of TF-IDF and n-grams approaches. As a result, key specific skillsets for various occupations are found. The proposed algorithm allows us to identify and standardize key skills which might be applicable to create a system of Russian classification for occupations and skills. In addition, the algorithm allows us to provide lists of the key combinations of skills that are in high demand among companies inside each particular occupation. |
Mathematical methods and algorithms of business informatics
|
84–92
|
In applied problems of business informatics related to data analysis (in particular, in the analysis and forecasting of time series, in the study of log files of business processes, etc.), problems of qualitative analysis arise. Qualitative analysis methods often use symbolic coding as a way of presenting information about the processes under study. In a number of situations, due to the fragmentation of such descriptions, the problem arises of reconstructing a complete symbolic description of a process (word) from its successive fragments (subwords). From the multiset of all subwords of a sufficiently large length, the original word is uniquely restored. In the case of insufficiently long subwords, several different reconstructions of the original word are possible. The number of feasible reconstructions can be reduced by determining the suffix and prefix of the reconstructed word. A method is proposed for determining the prefix and suffix of a word consisting of symbols each on the basis of multiset of subwords of a fixed length equal to. We accept the hypothesis that this multiset is generated by a window of a fixed length of one symbol shift in an unknown word. The method for determining the prefix and suffix is based on the construction and analysis of the matrix formed by subwords from written in rows in arbitrary order and the use of the operator acting on multisets of characters of the alphabet formed by neighboring columns of this matrix. The method is capable of determining the prefix and suffix, if for any from 1 to. If in the prefix and suffix only for some values of i, the characters in the corresponding positions are determined, and for the remaining characters. In the worst case, the method concludes that for any from 1 to, but does not determine the characters themselves. This is a situation in which the prefix and suffix coincide but cannot be determined. |
|
|