基于分散图决策的ε占优多目标粒子群算法施罗德声扩散体设计

发布时间：2018-04-23 15:16

  本文选题：声音扩散体 + 多目标粒子群　； 参考：《科学技术与工程》2017年32期

【摘要】：为提高声音扩散体设计的合理性降低算法计算复杂度,提出基于分散图决策的ε占优多目标粒子群算法(ε-MOPSO)的声音扩散体设计方法。首先,利用夫琅禾费理论建立施罗德扩散的声音扩散特性量化方法,获得1/3倍频带的扩散系数;并利用归一化方式消除极限尺寸下产生的边缘衍射散射效应。其次,构建声音扩散体多目标优化模型,通过对扩散系数进行重设置,消除扩散体重复和等价问题。然后采用ε-MOPSO算法,将目标空间分割成固定数量的n个网格,以保持种群解的多样性,实现声音扩散体参数优化;并采用分散图决策方式实现最终散体设计方案选择。最后,通过仿真对3种不同的设计模型进行了评价和选择。
[Abstract]:In order to improve the rationality of sound diffuser design and reduce the computational complexity of the algorithm, a method of sound diffusion volume design based on 蔚 -MOPSO-based 蔚 -dominated multi-objective particle swarm optimization (蔚 -MOPSO) is proposed. Firstly, by using Fraunhofer's theory, the sound diffusion characteristics of Schroeder diffusion are quantified, and the diffusion coefficient of 1 / 3 times frequency band is obtained, and the edge diffraction scattering effect under the limit size is eliminated by the normalized method. Secondly, the multi-objective optimization model of acoustic diffuser is constructed, and the problem of repetition and equivalence of diffuser is eliminated by re-setting the diffusion coefficient. Then 蔚 -MOPSO algorithm is used to divide the target space into a fixed number of n meshes in order to maintain the diversity of population solutions and optimize the parameters of acoustic diffuser. Finally, three different design models are evaluated and selected by simulation.
【作者单位】： 中国科学院武汉物理与数学研究所;
【基金】：国家自然科学基金(11404375) 国家重点基础研究发展计划(2012CB922101)资助
【分类号】：TP18
，

本文编号：1792508