第1期 李素，等：群智能算法优化支持向量机参数综述 。79· 3)大多数研究成果都在关注算法的前期后期 [11]AVCI E.Selecting of the optimal feature subset and kernel 收敛性问题，少部分研究中会将算法的收敛性和寻 parameters in digital modulation classification by using hy- 优精度结合在一起同时考虑，进行改进。群智能算 brid genetic algorithm-support vector machines:HGAS- 法在对SVM进行参数优化时，不能确保每种条件 VM[J].Expert systems with applications,2009,36(2): 下都拥有较强的寻优能力，同时也不能保证得到的 1391-1402. 最优参数对每种模型都拥有很好的分类和预测能 [12]HUANG C L.WANG C J.A GA-based feature selection 力。改进后的算法虽然可以同时具备寻优速度快、 and parameters optimizationfor support vector machines[J]. 寻优精度高、收敛到全局最优解、避免过早陷入局 Expert systems with applications,2006,31(2):231-240. 部最优等优点，但还需要对算法进行不断地研究与 [13]王琼瑶，何友全，彭小玲.基于改进遗传算法的支持向量 改进。 机参数优化方法U.计算机与现代化，2015(3)33-36. WANG Qiongyao,HE Youquan,PENG Xiaoling.Para- 参考文献： meters optimization of support vector machine based on [1]VAPNIK V N.Statistical learning theory[M].New York: improved genetic algorithm[J].Computer and moderniza- Wiley Press,1999. tion,2015(3:33-36. [2]POOLE A,KOTSIALOS A.Swarm intelligence algorithms [14]孟滔，周新志，雷印杰.基于自适应遗传算法的SVM参 for macroscopic traffic flow model validation with automat- 数优化.计算机测量与控制，2016,24(9)：215-217， ic assignment of fundamental diagrams[J].Applied soft 223 computing,2016.38:134-150. MENG Tao,ZHOU Xinzhi,LEI Yinjie.A parameters op- [3]DE SA A O,NEDJAH N,DE MACEDO MOURELLE L. timization method for an SVM based on adaptive genetic Distributed efficient localization in swarm robotic systems algorithm[J].Computer measurement and control,2016. using swarm intelligence algorithms[J].Neurocomputing, 24(9y:215-217.223. 2016.172:322-336. [15]高雷阜，张秀丽，佟盼.GASJ在SVM核参数优化中的 4]RAMADAN H S.BENDARY A F,NAGY S.Particle 应用).计算机工程与应用，2015,51(4)：110-114 swarm optimization algorithm for capacitor allocation prob- GAO Leifu,ZHANG Xiuli,TONG Pan.Application of lem in distribution systems with wind turbine generators[J]. GA SJ in SVM parameter optimization[J].Computer en- International journal of electrical power and energy systems, gineering and applications,2015,51(4):110-114 2017,84:143-152. [16]SAJAN K S,KUMAR V,TYAGI B.Genetic algorithm [5]KEERTHI SS.Efficient tuning of SVM hyperparameters based support vector machine for on-line voltage stability using radius/margin bound and iterative algorithms[J].IEEE monitoring[J].International journal of electrical power and transactions on neural networks,2002,13(5):1225-1229. energy systems,2015,73:200-208. 6]CHAPELLE O,VAPNIK V.BOUSQUET O,et al.Choos- [17]CHOU J S,CHENG Minyuan,WU Yuwei,et al.Optimiz- ing multiple parameters for support vector machines[J].Ma- ing parameters of support vector machine using fast messy chine learning,2002,46(1/2/3):131-159. [7]HOLLAND J H.Adaptation in natural and artificial sys- genetic algorithm for dispute classification[J].Expert sys- tems[Ml.2nd ed.Cambridge:MIT Press,1992. tems with applications,2014,41(8):3955-3964 [8]吴君，张京娟.采用遗传算法的多机自由飞行冲突解脱策 [18]DUAN Li,GUO Long,LIU Ke,et al.Characterization and 略U.智能系统学报，2013,8(1)：16-20 classification of seven Citrus herbs by liquid chromato- WU Jun,ZHANG Jingjuan.Conflict resolution of multiple graphy-quadrupole time-of-flight mass spectrometry and airplanes in free flight based on the genetic algorithm[J]. genetic algorithm optimized support vector machines[J] CAAI transactions on intelligent systems,2013,8(1):16- Journal of chromatography A,2014,1339:118-127 20 [19]DORIGO M,GAMBARDELLA L M.Ant colony system: [9]DIABAT A.DESKOORES R.A hybrid genetic algorithm a cooperative learning approach to the traveling salesman based heuristic for an integrated supply chain problem[J]. problem[J].IEEE transactions on evolutionary computa- Journal of manufacturing systems,2016,38:172-180. tion,1997,1(1):53-66. [10]CHEN Peng,YUAN Lifen,HE Yigang,et al.An im- [20]DORIGO M,STUTZLE T.Ant colony optimization[M]. proved SVM classifier based on double chains quantum ge- Cambridge:MIT Press,2004. netic algorithm and its application in analogue circuit dia- [21]裴振兵，陈雪波.改进蚁群算法及其在机器人避障中的 gnosis[J].Neurocomputing,2016,211:202-211. 应用).智能系统学报，2015,10(1)上90-96.3) 大多数研究成果都在关注算法的前期后期 收敛性问题，少部分研究中会将算法的收敛性和寻 优精度结合在一起同时考虑，进行改进。群智能算 法在对 SVM 进行参数优化时，不能确保每种条件 下都拥有较强的寻优能力，同时也不能保证得到的 最优参数对每种模型都拥有很好的分类和预测能 力。改进后的算法虽然可以同时具备寻优速度快、 寻优精度高、收敛到全局最优解、避免过早陷入局 部最优等优点，但还需要对算法进行不断地研究与 改进。 参考文献： VAPNIK V N. Statistical learning theory[M]. New York: Wiley Press, 1999. [1] POOLE A, KOTSIALOS A. Swarm intelligence algorithms for macroscopic traffic flow model validation with automat￾ic assignment of fundamental diagrams[J]. Applied soft computing, 2016, 38: 134–150. [2] DE SÁ A O, NEDJAH N, DE MACEDO MOURELLE L. Distributed efficient localization in swarm robotic systems using swarm intelligence algorithms[J]. Neurocomputing, 2016, 172: 322–336. [3] RAMADAN H S, BENDARY A F, NAGY S. Particle swarm optimization algorithm for capacitor allocation prob￾lem in distribution systems with wind turbine generators[J]. International journal of electrical power and energy systems, 2017, 84: 143–152. [4] KEERTHI S S. Efficient tuning of SVM hyperparameters using radius/margin bound and iterative algorithms[J]. IEEE transactions on neural networks, 2002, 13(5): 1225–1229. [5] CHAPELLE O, VAPNIK V, BOUSQUET O, et al. Choos￾ing multiple parameters for support vector machines[J]. Ma￾chine learning, 2002, 46(1/2/3): 131–159. [6] HOLLAND J H. Adaptation in natural and artificial sys￾tems[M]. 2nd ed. Cambridge: MIT Press, 1992. [7] 吴君, 张京娟. 采用遗传算法的多机自由飞行冲突解脱策 略[J]. 智能系统学报, 2013, 8(1): 16–20. WU Jun, ZHANG Jingjuan. Conflict resolution of multiple airplanes in free flight based on the genetic algorithm[J]. CAAI transactions on intelligent systems, 2013, 8(1): 16– 20. [8] DIABAT A, DESKOORES R. A hybrid genetic algorithm based heuristic for an integrated supply chain problem[J]. Journal of manufacturing systems, 2016, 38: 172–180. [9] CHEN Peng, YUAN Lifen, HE Yigang, et al. An im￾proved SVM classifier based on double chains quantum ge￾netic algorithm and its application in analogue circuit dia￾gnosis[J]. Neurocomputing, 2016, 211: 202–211. [10] AVCI E. Selecting of the optimal feature subset and kernel parameters in digital modulation classification by using hy￾brid genetic algorithm-support vector machines: HGAS￾VM[J]. Expert systems with applications, 2009, 36(2): 1391–1402. [11] HUANG C L, WANG C J. A GA-based feature selection and parameters optimizationfor support vector machines[J]. Expert systems with applications, 2006, 31(2): 231–240. [12] 王琼瑶, 何友全, 彭小玲. 基于改进遗传算法的支持向量 机参数优化方法[J]. 计算机与现代化, 2015(3): 33–36. WANG Qiongyao, HE Youquan, PENG Xiaoling. Para￾meters optimization of support vector machine based on improved genetic algorithm[J]. Computer and moderniza￾tion, 2015(3): 33–36. [13] 孟滔, 周新志, 雷印杰. 基于自适应遗传算法的 SVM 参 数优化[J]. 计算机测量与控制, 2016, 24(9): 215–217, 223. MENG Tao, ZHOU Xinzhi, LEI Yinjie. A parameters op￾timization method for an SVM based on adaptive genetic algorithm[J]. Computer measurement and control, 2016, 24(9): 215–217, 223. [14] 高雷阜, 张秀丽, 佟盼. GA_SJ 在 SVM 核参数优化中的 应用[J]. 计算机工程与应用, 2015, 51(4): 110–114. GAO Leifu, ZHANG Xiuli, TONG Pan. Application of GA_SJ in SVM parameter optimization[J]. Computer en￾gineering and applications, 2015, 51(4): 110–114. [15] SAJAN K S, KUMAR V, TYAGI B. Genetic algorithm based support vector machine for on-line voltage stability monitoring[J]. International journal of electrical power and energy systems, 2015, 73: 200–208. [16] CHOU J S, CHENG Minyuan, WU Yuwei, et al. Optimiz￾ing parameters of support vector machine using fast messy genetic algorithm for dispute classification[J]. Expert sys￾tems with applications, 2014, 41(8): 3955–3964. [17] DUAN Li, GUO Long, LIU Ke, et al. Characterization and classification of seven Citrus herbs by liquid chromato￾graphy–quadrupole time-of-flight mass spectrometry and genetic algorithm optimized support vector machines[J]. Journal of chromatography A, 2014, 1339: 118–127. [18] DORIGO M, GAMBARDELLA L M. Ant colony system: a cooperative learning approach to the traveling salesman problem[J]. IEEE transactions on evolutionary computa￾tion, 1997, 1(1): 53–66. [19] DORIGO M, STÜTZLE T. Ant colony optimization[M]. Cambridge: MIT Press, 2004. [20] 裴振兵, 陈雪波. 改进蚁群算法及其在机器人避障中的 应用[J]. 智能系统学报, 2015, 10(1): 90–96. [21] 第 1 期 李素，等：群智能算法优化支持向量机参数综述 ·79·