[1]杨春玲,王来来,朱 敏.采用改进的粒子群算法训练CNNE模型[J].智能系统学报,2007,2(3):67-72.
YANG Chun-ling,WANG Jian-lai,ZHU Min.Using the improved particle swarm optimization to train the CNNE model[J].CAAI Transactions on Intelligent Systems,2007,2(3):67-72.
点击复制
YANG Chun-ling,WANG Jian-lai,ZHU Min.Using the improved particle swarm optimization to train the CNNE model[J].CAAI Transactions on Intelligent Systems,2007,2(3):67-72.
采用改进的粒子群算法训练CNNE模型
《智能系统学报》[ISSN 1673-4785/CN 23-1538/TP] 卷:
2
期数:
2007年第3期
页码:
67-72
栏目:
学术论文—机器学习
出版日期:
2007-06-25
- Title:
- Using the improved particle swarm optimization to train the CNNE model
- 文章编号:
- 1673-4785(2007)03-0074-06
- Keywords:
- CNNE model; particle swarm optimization; gradient
- 分类号:
- TP183
- 文献标志码:
- A
- 摘要:
- 提出用人工智能算法——粒子群优化算法(PSO)对CNNE模型进行训练，并针对标准粒子群算法易限于局部极小点的局限性，采用了一种带有梯度加速的粒子群算法，通过引入梯度信息来影响粒子速度的更新.为防止陷入局部最优，在群体最优信息陷入停滞时，对部分粒子进行重新初始化，从而保持群体的活性，减小群体陷入局优的可能性.采用粒子群算法训练的CNNE模型较原来的分布式最速下降法而言，在保证精度的前提下，提高了算法的收敛速度，解决了发射率的在线实时测量问题.
- Abstract:
- An artificial intelligence algorithm - particle swarm optimization (PSO) were proposed to train the CNNE model. Aiming at the limitation of the sta ndard particle swarm optimization can be easily restricted in the local optimum point, a kind of particle swarm optimization(PSO)algorithm with gradient acceler ation is adopted by adding gradient information to influence the update of veloc ities of the particles. When the optimum information of the swarm is stagnant, s ome particles in the population are initialized again to reduce the possibility of trapping in local optimum. Comparing with the step steepest descent algorithm , using the particle swarm optimization algorithm to train the CNNE model improv es the speed of convergence of the algorithm on the premise of keeping precision , which solves the online realtime measurement of emissivity.
- 参考文献/References:
-
［1］YANG Chunling, YU Yong, ZHAO Dongyang, et al. Study on modeling of m ultispectral emissivity and optimization algorithm［J］. IEEE Trans.on Neural Ne tworks, 2006,17(1): 238-242.
［2］YANG Chunling, ZHAO Dongyang ， DAI Jingming. A new method for constructin g spectral emissivity models for measuring the real temperature of targets［J］. IEEE Trans. on Instrumentation and Measurement, 2005,54(6): 2549-2553.
［3］EBERHART R C, KENNEDY J. A new optimizer using particl swarm theory［A］. Proceedings of Sixth International Symposium Micro Machine and Human Science［C ］. Nagoya, Japan, 1995.
［4］EBERHART R C, KENNEDY J. Swarrn intelligence［M］. San Francisco: Mornan K aufmann Publishers, 2001.
［5］KENNEDY J, EBERHART R C. Particle swarm optimization［A］. Proceedings of IEEE International Conference on Neutral Networks［C.］Perth, Australia, 1995.
［6］CLERE M, KENNEDY J. The particle swarm-explosion, stability and convergenc e in a multidimensional complex space［J］. IEEE Trans.on Evolutionary Computati on, 2002,6(1): 58-73.
［7］BERGH F V D. An analysis of particle swarm optimizer［D］. PhD thesis. Dep artment of Computer Science, University of Pretoria, South Africa, 2002
［8］王俊伟,汪定伟. 一种带有梯度加速的粒子群算法［J］.控制与决策. 2004,19 (11): 1298-1230.
?WANG Junwei, WANG Dingwei. Particle swarm optimization algorithm with gradient a cceleration［J］. Control and Decision， 2004,19(11): 1298-1230.
- 相似文献/References:
-
[1]刘宏达,马忠丽.均匀粒子群算法[J].智能系统学报,2010,5(4):336.
LIU Hong-da,MA Zhong-li.A particle swarm optimization algorithm based on uniform design[J].CAAI Transactions on Intelligent Systems,2010,5():336.
[2]贾旋,周治平.一种基于模板缩减的新型粒子群遗传聚类算法[J].智能系统学报,2016,11(4):561.[doi:10.11992/tis.201507026]
JIA Xuan,ZHOU Zhiping.A novel PSO-GGA for clustering based on pattern reduction[J].CAAI Transactions on Intelligent Systems,2016,11():561.[doi:10.11992/tis.201507026]
[3]陈强,王宇嘉,梁海娜,等.目标空间映射策略的高维多目标粒子群优化算法[J].智能系统学报,2021,16(2):362.[doi:10.11992/tis.202006042]
CHEN Qiang,WANG Yujia,LIANG Haina,et al.Multi-objective particle swarm optimization algorithm based on an objective space papping strategy[J].CAAI Transactions on Intelligent Systems,2021,16():362.[doi:10.11992/tis.202006042]
备注/Memo
收稿日期：2006-10-13.
基金项目：
国家自然科学基金资助项目（60377037）；
黑龙江省自然科学基金资助项目（F200413）.
作者简介：
杨春玲，女，1965年生，教授，博士生导师， IEEE会员；东北地区高校电子技术研究会理事；主要研究方向为神经网络建模及优化算法、目标红外辐射特性参数的测量、红外图像处理、微弱信号检测及FPGA硬件实现技术等. 在国内外发表论文40余篇，被SCI检索5篇，主编教材4本.承担省部级科研项目6项.
E-mail：yangcl1@hit.edu.cn.
王来来，男，1982年生，硕士研究生，主要研究方向为基于FPGA/CPLD的数字系统设计，人工神经网络建模等.
朱 敏，男，1981年生，主要研究方向为红外检测.
更新日期/Last Update:
2009-05-07