Search academic texts

Information × Registration Number 0212U004704, 0111U007500 , R & D reports Title Development of algorithms of the fast global optimization for the electromagnetics synthesis popup.stage_title Head Kirilenko A. A., Registration Date 23-01-2012 Organization Institute of Radiophisics and Electronics NASU popup.description2 The subject of the research is the methods and global optimization algorithms, namely genetic algorithm (GA) and the algorithm based on the particle swarm method (PSM), which are the most popular tools for electromagnetic synthesis. The project aims are to develop new efficient algorithms for global optimization based on GA concepts and PSM able to quickly find optimal solutions in complex and many parametric problems, typical for electromagnetic synthesis. The importance and relevance of this work are explained by the constantly increasing requirements for electromagnetic devices and systems design where traditional methods are very difficult because of excessive requirements to computer time and resources. Presentation of new more efficient algorithms will reduce time for development of new devices (antennas and waveguide structures) and to improve their performance. Two of the most popular concepts of global optimization are ыtudied and implemented, namely genetic algorithm (GA) and algorithm based on the particle swarm method (PSM). The method for adapting of the parameters of algorithms to ensure optimal allocation of effort on a global and local search. A few schemes of the physical parameters encode to reduce its dimensionality and computational complexity are proposed. The possibility of combining different global and local algorithms for optimization problem under different conditions (type of objective function, type of variables, etc.) are studied. A hybrid algorithm based on GA capable to optimize functions with many local extrema is built. An adaptive algorithm based on PSM rapidly optimizing different types of functions under conditions of uncertainty region of permissible values for optimization is developed. Product Description popup.authors Бліннікова Ольга Анатоліївна Приколотін Сергій Анатолійович Стешенко Сергій Олександрович popup.nrat_date 2020-04-02 Close