The object of research - the management process in distributed transport logistics systems in conditions of uncertainty.
The subject of research - models and methods of traffic management in conditions of uncertainty
The purpose of research is to increase the management efficiency in the distributed system of transport logistics in conditions of uncertainty. This goal is achieved by solving the following tasks:
- development of a method for constructing effective routes taking into account a set of criteria; - development of an exact method of forming a transportation plan in the system "suppliers - consumers" by the criterion - probability of not exceeding the random value of the total cost of transportation of the allowable threshold; - development a construction method of a compromise route with the using concessions in relation to the optimal plan according to the main criterion; - development of a method for calculating the capacity of intermediate points in the extensive transport network in conditions of uncertainty; - development of a method for optimizing the structure of the distributed transport network when calculating distances in the metric "urban quarters"; - development of an exact method for solving the problem of fractionalquadratic programming ;
- development of a fast approximate method for solving nonlinear optimization problems under conditions of uncertainty;
Research methods. Probability theory and fuzzy mathematics in the development of models of transport logistics systems that describe the processes of their operation in conditions of uncertainty.
Methods of solving optimization problems in theoretical-probabilistic and fuzzy formulations. The method of continuous linear programming for problems whose initial data are given by their modal values. In the dissertation research with the use of modern mathematical models and methods of probability theory, random processes, fuzzy set theory, tools of optimization methods and continuous linear programming set and solved an important scientific and technical problem of control in a distributed transport system in fuzzy initial data. The practical significance of the obtained results is to develop a set of models and management methods in distributed transport logistics systems in conditions of uncertainty about the values of the original data. The developed models and methods together solve an important scientific and applied problem of complex systems management and create a theoretical foundation for solving practical problems. Developed methods of forming optimal routes in the system "suppliers-deliveryconsumers" provide the opportunity to significantly increase the efficiency of transportation in solving real problems of transport logistics.
Management tasks in distributed transport systems are an important element of a wide range of problems solved by the methods of general logistics theory. The corresponding transport issue combines a set of issues with the same type of mathematical model, which are solved by linear programming methods. In the canonical formulation, the task is to find a plan of transportation of some homogeneous product from many suppliers to many consumers, optimal in terms of some selected criterion. To solve the problem, traditionally, the following two approaches are used.
The first - it is assumed that its parameters are set by their deterministic values. Second - random parameters of the problem are replaced by their average values. It is clear that the adequacy of the relevant models is not satisfactory. This research develops methods for solving transportation management problems taking into account the real uncertainty of the original data.
A large number of foreign and domestic scientists have been working on the development of methods for solving transport problems and their various modifications: J.Gass, Т.С.Motzkin, K.B.Halley, A.A.Corban, M.Cerсhes, B.G.Dantzig, C.Mihu, M.Vlach, J.Moravec, G.Smith, D.B. Yudin, E.G.Goldshtain, B.S. Verhovskiy, V.А. Yemelyanov, V.V. Ivanov, Y.М. Nerush, М.P. Gordon, V.А. Stahanov, V.S. Lukinskuy and others. In the works of these authors, linear models and methods for solving classical two-index transport problems in a deterministic formulation are exhaustively considered. However, this very superficially affects the problem of the possible high dimensionality of such tasks, as well as the features and difficulties of their solution in cases of uncertainty of the source data. The need and importance of considering and solving these problems is constantly growing. In this regard, task of the dissertation devoted to models and methods of solving problems of transport logistics in the conditions of real uncertainty of the initial data, is relevant.
