Research on distributed collaborative optimization technology

Complex engineering optimization problems, such as design optimization of complex systems, are difficult to solve due to large-scale computation and complex simulation in order to compute a function value, so the distributed computing technology based on decomposition-coordination theory is very attention -getting to design engineers. Considering some disadvantages of the present computing methods and frames, this paper puts forward a distributed computing frame, Collaborative Subspace Optimization Frame (CSOF). Then, the paper delves into the coordination problem of distributed collaborative optimization systems for dealing with the conflict among related variables in sub-optimization problems. A complete coordination model is built, in which the coordination process is treated as an optimization problem itself And a genetic algorithm based coordination strategy of sub-optimization is studied. The collaborative optimization system is implemented using multi-agent technology and the design and implementation of optimization agents are discussed in detail, and CORBA technology is used to implement communication between the components of the optimization agents. An example demonstrates the computing method efficient and the multi-agent system reliable and flexible.