Development of Control Techniques for Nonlinear Networks and Applications

点击翻译按钮获取中文摘要

1. Design, Control Planning and Real-Time Control of Water Distribution Networks:First, we have developed a general method of network design which enables us to judge a trade-off relationship between reliability and cost of water supply, and second, its modified version for extension of existing networks. In relation to these, we have proposed an idea of generalized steady-flow analysis and an effective algorithm to obtain its most meaninful solution. For control planning and real-time control problems, various remarkable improvements of the existing methods have been proposed from both mathematical and practical viewpoints. Many of our ideas have already been put into practical use with favorable comments.2. Analysis and Simulation of Oscillations in Nonlinear Circuits with Many-Degrees of Freedom,:We have investigated various oscillations, such as synchronizations and multi-mode oscillations, in van der Pol type oscillators with many-degrees of freedom, particularly those with strong … More nonlinearity, by means of the averaged potential, which was developed by our group earlier, and computer simulation. The results give us a solid foundation for designing an information processing machine by use of oscillators, and further a new viewpoint to the mechanism of almost-daily rhythm of life.3. Control and Management of Energy Supply Networks:Supply networks of energy, such as electricity and town-gas, are very much complicated ones. Several ideas for control and management of consumption loads in those networks have been developed by constructing models of nonlinear optimization, game and behavior analysis types.4. Self-Organization of Systems and Investigation of Associative Memory Networks with Connection Plasticity:We have considered an essential role of nonlinearity in the mechanism of self-organization of systems. A variety of new ideas have been proposed for formulating, analyzing and designing an autoassociative memory network in which connections among elements (neurons) are plastically variable depending upon signals flowing. Particularly, the idea of a unified formulation with the perceptron is quite useful, and trials for localization and nonlinearization of the connections (signal transformations) have yielded promising results for improvement of dynamic characteristics of the memory network. Less