| Summary: | 碩士 === 國立雲林科技大學 === 工業工程與管理研究所碩士班 === 101 === Nano-particles is an advanced material, the preparation process is very important, because of the preparation process will generate a lot of sensitivity in process of complexity and variability. At present, the industry usually uses wet-type mechanical milling process to make nano-particles, it can use the high-speed rotating of stirring blade to pulverize liquid of micro particle, and add a solvent and dispersing agent in order to avoid particles condensing. Due to the smaller grain size and the small variability of nano-particles, it can enhance its surface activity and catalytic, therefore, how to control the better nano-particles parameters will be a challenge.
The Particle Swarm Optimization (PSO) is an algorithm that derives from observing the foraging behavior of bird populations. It has the advantage of speed up convergence, but it also has the shortcoming of early convergence and fall into local optimum. Therefore, this thesis proposes a Revised Adaptive Particle Swarm Optimization (RAPSO). In addition to add Ren-De You(2012) the concept of random particles, also add the concept of particle mating, and adjust the parameters of inertia weight and learning factors by the variation of iterations and fitness values. Expectation of adding the modified strategy, it not only can preserve the advantage of Particle Swarm Optimization, but also can improve the shortcoming of Particle Swarm Optimization, and then enhance solution accuracy.
The research results showed that in the function test and nano-particles milling process parameter optimization problem, the Revised Adaptive Particle Swarm Optimization for solving capacity and performance can be improve and enhance.
|
|---|
