|
As a new evolution of computing technology, swarm intelligence has become a new hotspot. Completed theoretical and applied research swarm intelligence method is an effective solution to most global optimization problems. More importantly, the swarm intelligent the potential parallel and distributed characteristics for dealing with a large amount of data exists in the form of a database providing technical assurance. Compared with traditional optimization algorithms, particle swarm optimization in multi-dimensional function optimization, dynamic targeted hunt Excellence has a fast convergence and high quality solutions, robustness, particularly for mechanical engineering applications. The particle swarm algorithm exists in the later stage of evolution, the search speed is slow, easy to fall into local minima and search a long time to solution, and accuracy is not high, so the improved algorithm has become an essential issue. In this paper, the two most of the research: for decreasing inertia linear particle swarm algorithm can not adapt to the complex nonlinear optimization problem of the search process, two improved particle swarm algorithm. A dynamic inertia weight particle swarm optimization algorithm (DIPSO), the introduction of these two parameters of the evolution speed factor and aggregation factor. For the minimum of the optimization problem: the smaller the evolution speed factor, indicating that the speed of evolution of the sooner, the algorithm can continue to search in the search space can be reduced inertia weight value, making the particle swarm search space in a small area, in order to find the optimal solution faster. If the particle is more dispersed, the particles will not be easy to fall into local optimal solution, with the improvement in the level of aggregation of the particle swarm algorithm is easy to fall into local optimal solution, then, should be increased inertia weight, thus increasing the search space of the particle swarm to improve the global optimization ability of the particle swarm. So, the improved algorithm inertia weight can be expressed as a function of the evolution speed factor and aggregation factor. Each iteration of the algorithm can dynamically change based on the current evolution speed factor and aggregation degree factor inertia weight, so that the algorithm has a dynamic self-adaptive. An adaptive random inertia weight particle swarm optimization algorithm (ARIWPSO), the introduction of group fitness as inertia weight control parameters, the inertia weight with the change in the population's fitness, the adaptive inertia weight particle swarm algorithm. For an improved constrained single-objective particle swarm optimization algorithm, the optimal design of three instances of Mechanical Engineering. The results show that: the two improved particle swarm optimization algorithm-based optimization design is feasible, provide new ideas and methods for complex mechanical optimization design.
|