Search academic texts

Information × Registration Number 0219U102296, 0114U006580 , R & D reports Title The development of theoretical foundations, methods and tools for designing local and regional networks monitoring air quality popup.stage_title Head Artemchuk Volodymyr O., Registration Date 26-12-2019 Organization Pukhov Institute for Modelling in Energy Engineering National Academy of Sciences of Ukraine popup.description2  It is determined that the most difficult for the study is the case when the task of optimizing the network of atmospheric air monitoring (AAM) is: multicriteria; the optimization area is not convex or disconnected; the number of sensor nodes in the ISAP network is unknown a priori; optimizes the mixed (hybrid) network of AAM in the presence of an existing network, part of which can be transferred and part - not; there are several limitations in the problem, that is, it is a conditional optimization problem; constraints or criteria are non-linear, non-convex, etc., that is, the problem of nonlinear non-convex programming, etc. In this case, finding the optimal solution can only be found by the complete search method. However, in this case, the full batch method, because of its large computational complexity, can be applied only to small territories (cities), so there is a need to use heuristic optimization algorithms. The following methods and algorithms have been considered: evolutionary search methods (including genetic algorithms), greedy algorithms, ant algorithms, Drop algorithm, Maranzana algorithm, Teitz-Bart algorithm or Bump algorithm, a GRIA (Global-Regional Interchange Approach) algorithm. Evolutionary search methods and greedy algorithms were chosen for further research. In general, to solve this problem it is proposed to use a greedy algorithm, which, in comparison with the methods of evolutionary search, provides a faster finding of the closest to the optimal or optimal solution. Two algorithms have been developed for solving this problem, based on different ways of combining the greedy algorithm with the method of full search, which allowed to improve the accuracy of the solutions obtained by the classical greedy algorithm. It is substantiated that the development of AAM network design software is appropriate within the AISEEM system. A number of relevant structural and information models in the form of UML diagrams have been developed. AISEEM has dev Product Description popup.authors Bugaev Oleksandr P. Kameneva Iryna P. Yatsyshyn Andrii V. popup.nrat_date 2020-04-02 Close