PSO-based Multi-objective Optimization Algorithm Research And Its Applications

After nearly a century of research and development, optimization theory and related algorithms have increasingly come into mature. Traditional optimization algorithms have been widely used a wide range of industrial process due to their complete theory , high efficiency and robustness. However, the traditional strategy of optimizing calculations based on the gradient of the function of continuity, derivative of the presence of a relatively high demand, thus limiting its further development. Moreover, based on the gradient optimization strategy often used to initialize points more sensitive, a good initialization of the algorithm itself would lead to easy solutions faster, and a relatively poor rate of basically can make itself greatly reduced, or even no guarantee that the algorithms themselves to the best. From the 1960s onwards, to GA to represent the beginning of wisdom algorithms thrived, and its main features in a random search strategy can not rely on the selection and do not consider whether the function itself is continuous or differential.The main aim of the study is the convergence of dynamics model of Particle Swarm Optimization(PSO), and so on, through statistical analysis of the probability , the convergence precision and speed is discussed . Subsequently, the idea of a multi-objective optimization algorithms based on PSO is presented , and algorithm efficiency and the other fellow made a comparative analysis algorithms, the algorithms have proven their relative rationality and superiority. Finally, the paper introduced the PSO and PSO based on the multi-objective optimization algorithms in a specific application for the actual project.The main work includes :1) The most common algorithms for the simplified dynamics model of PSO. Firstly , the model was discussed based on Linear system theory and presented the sufficient condition of the model if convergence. Then , the model was derived based on interval systems theory and linear matrix inequality . Therefore, based on the random process theory, the model of PSO was derived and proved sufficient condition.2) present a multi-objective optimization algorithms based on PSO strategy, and through two sets are combined and repeated dynamic adjustments, resulting in algorithm efficiency and accuracy reached a better balance .Function testing and test results show that the algorithm is relatively better effective and rapid performance standards in the comparison strategy to demonstrate the algorithm than some similar...