ID 原文 译文
7104 针对现有量子粒子群优化算法的多参数(≥5)优化问题易收敛到局部最优解、且无法判定优化结果全局性的问题,提出了带全局判据的改进量子粒子群优化算法。 Quantum particle swarm optimization algorithm in view of the existing multi-parameter optimization (5) or higher easy convergence to the local optimal solution, and unable to determine the problem of global optimization results, puts forward the improved quantum particle swarm optimization algorithm with global criterion.
7105 在惯性权重自适应调整的量子粒子群优化算法基础上,进行了粒子位置周期性变异,以及随粒子进化速度和聚集度变化的搜索范围变异。 In adaptive inertia weight adjustment based on quantum particle swarm optimization algorithm, the particle position periodic variation, as well as with the evolution speed and concentration of particles change the search range of variation.
7106 依据粒子聚集度大小,建立了判定优化结果全局性的全局收敛判据。 According to the size of particle concentration, established the global convergence criterion for global optimization results.
7107 以典型标准函数和乘波体外形多参数优化问题为算例,验证了改进算法和全局判据的可靠性。 In a typical standard function and waverider shape multi-parameter optimization problem for example, the reliability of the improved algorithm and global criterion.
7108 结果表明,改进算法的全局搜索能力明显提高,优化结果真实可靠,全局判据实用性强。 The results showed that the improved algorithm obviously improve the global search ability, optimized result is reliable, global criterion of strong practicability.
7109 针对复杂仿真模型验证中海量数据的相似性分析问题,提出了一种基于集成学习的仿真模型验证方法。 In view of the complex simulation model to verify the similarity analysis of huge amounts of data in the question, proposed based on an integrated study of the simulation model validation method.
7110 将仿真时间序列与参考时间序列的相似性分析问题转换为相似性等级分类问题。 Simulation time series with reference time series similarity analysis problem is converted into a similarity classification problem.
7111 进而利用神经网络、支持向量机、集成学习等机器学习方法,设计了一种集成分类系统对时间序列的相似性等级进行分类。 And then use neural network and support vector machine (SVM), integrated learning machine learning method, we design a integrated classification system for time series similarity level classification.
7112 为了增强基分类器的多样性,提出了基于惩罚因子的多样性筛选准则; In order to enhance the diversity of the base classifier, is proposed based on the diversity of punishment factor screening criteria;
7113 通过挑选具有最大差异性的基分类器,构造高性能集成分类系统。 By selecting base classifier with the largest genetic diversity, structure of high-performance integrated classification system.