[1]张 永,吴晓蓓,徐志良,等.基于Pareto多目标遗传算法的模糊系统设计[J].南京理工大学学报(自然科学版),2007,(04):430-434.
 ZHANG Yong,WU Xiao-bei,XU Zhi-liang,et al.Design of Fuzzy Systems Based on Pareto Multi-objective Genetic Algorithm[J].Journal of Nanjing University of Science and Technology,2007,(04):430-434.
点击复制

基于Pareto多目标遗传算法的模糊系统设计
分享到:

《南京理工大学学报》(自然科学版)[ISSN:1005-9830/CN:32-1397/N]

卷:
期数:
2007年04期
页码:
430-434
栏目:
出版日期:
2007-08-30

文章信息/Info

Title:
Design of Fuzzy Systems Based on Pareto Multi-objective Genetic Algorithm
作者:
张 永 吴晓蓓 徐志良 黄 成 胡维礼
南京理工大学自动化学院, 江苏南京210094
Author(s):
ZHANG YongWU Xiao-beiXU Zhi-liangHUANG ChenHU Wei-li
School of Automation,NUST,Nanjing 210094,China
关键词:
TS模糊模型 模糊分类系统 模糊聚类 遗传算法 解释性
Keywords:
A novel approach to construct accurate and in terpretable fuzzy system s based on Pareto mult-i objective genet ic algor ithm is proposed. The approach is composed of tw o phases: the in it ial fuzzy system is identif ied using fuzzy clustering algor ithm the Pittsburgh-sty le rea-l coded genetic a-l gorithm is used to optim ize the structure and parameters o f the fuzzy systems and the three-objective function based on NSGA-Ⅱ a lgorithm comb ines the interpretab ility indices and the precision index. In order to improve the interpretab ility of the fuzzy system the sim ilarity-dr iven rule base simp lificat ion techniques are used to reduce the fuzzy system. The proposed approach is applied to several benchmark prob lems and the results show its va lid ity.
分类号:
TP18
摘要:
提出一种基于Pareto多目标遗传算法生成一组精确性和解释性较好折衷模糊系统的方法。该方法采用模糊聚类算法辨识初始的模糊模型,利用匹茨堡型实数编码的遗传算法对初始模糊模型的结构和参数进行优化,基于NSGA-Ⅱ算法的目标函数同时考虑模型的精确性和解释性;最后,在算法中利用基于相似性的模型简化方法约简模糊系统。利用该方法对两个Benchmark系统进行建模,仿真结果验证了该方法的有效性。
Abstract:
A novel approach to construct accurate and in terpretable fuzzy system s based on Pareto mult-i objective genet ic algor ithm is proposed. The approach is composed of tw o phases: the in it ial fuzzy system is identif ied using fuzzy clustering algor ithm; the Pittsburgh-sty le rea-l coded genetic a-l gorithm is used to optim ize the structure and parameters o f the fuzzy systems, and the three-objective function based on NSGA-Ⅱ a lgorithm comb ines the interpretab ility indices and the precision index. In order to improve the interpretab ility of the fuzzy system, the sim ilarity-dr iven rule base simp lificat ion techniques are used to reduce the fuzzy system. The proposed approach is applied to several benchmark prob lems, and the results show its va lid ity.

参考文献/References:

[ 1] M iku t R, Jake l J, Gro ll L. In terpretab ility issues in data-based lea rning o f fuzzy system s [ J]. Fuzzy Sets and System s, 2005, 150 ( 2): 179- 197.
[ 2] Roubo sH, SetnesM. Com pact and transparent fuzzy models and class ifiers through ite rative comp lex ity reduction [ J]. IEEE Trans on Fuzzy Sy stem s, 2001, 9( 4): 516- 524.
[ 3] DelgadoM R, Zuben F V, Gom ide F. Mult-i objective dec ision making: Towards mi provem ent of accuracy, interpretability and design autonomy in hierarchica l genetic fuzzy systems [A]. Proc o f IEEE Int Conf on Fuzzy System s[ C]. Hono lulu, Hawai:i [ s. n. ], 2002. 1 222- 1 227.
[ 4] Ho S-Y, Chen H-M. Design o f accu rate c lassifiers w ith a com pac t fuzzy- rule base using an evo lutionary scatte r partition o f fea ture space [ J]. IEEE T rans on System s, M an and Cybernetics ( Pa rt B ), 2004, 34 ( 2): 1 031- 1 044.
[ 5] Guillaum e S. Designing fuzzy in ference system s from data: An in terpretab ility-or iented rev iew [ J] . IEEE Trans on Fuzzy System s, 2001, 9 ( 3): 426- 443.
[ 6] Jin Y. Advanced fuzzy system s design and applications [M ] . New Yo rk: Phys ica-l ve r,l 2003.
[ 7] X ing Z-Y, J ia L-M, Zhang Yong. A case study of data- dr iven inte rpre tab le fuzzy m odeling [ J]. ACTA Autom atica S in ica, 2005, 31 ( 6): 815- 824.
[ 8] 张永, 吴晓蓓, 向峥嵘, 等. 基于多目标进化算法的 高维模糊分类系统的设计[ J]. 系统仿真学报, 2007, 19( 1) : 210- 214. [ 9 ] Setnes M, Babuska R, Kaym ak U, et a.l Smi ilarity m easures in fuzzy rule base smi plification [ J]. IEEE Trans on System s, Man and Cybernetics, 1998, 28 ( 3): 376- 386.
[ 10] Rui P P, Anton io D. Interpretab ility and learn ing in neuro-fuzzy system s [ J] . Fuzzy Sets and System s, 2004, 147 ( 1): 17- 38.
[ 11] Deb K, Pratap A, Agarw al S, M eyar ivan T. A fast and e litist mu ltiobjective genetic a lgor ithm: NSGA- II [ J]. IEEE Trans Evolutionary Computation, 2002, 6 ( 2): 182- 197.
[ 12] JinW Y, von Seelen, Sendhoff B. On generating FC3 fuzzy ru le system s from data using evo lu tion strateg ies [ J] . IEEE T rans System s, M an Cyberne tics, 1999, 29 ( 6): 829- 845.
[ 13] Nauck D, Kruse R. Neuro- fuzzy sy stem s for function approx ima tion [ J] . Fuzzy Sets and Sy stem s, 1999, 101 ( 2): 261- 271.
[ 14] W ang J S, Lee G. C S. Se l-f adaptive neuro- fuzzy inference system for c lassification app lication [ J]. IEEE Trans Fuzzy System s, 2002, 10 ( 6): 790- 802. [ 15 ] Russo M. Genetic fuzzy learn ing [ J]. IEEE Trans Evo lu tionary Computation, 2000, 4 ( 3) : 259- 273.

备注/Memo

备注/Memo:
基金项目: 国家自然科学基金( 60474034)
作者简介: 张永( 1969 - ) , 男, 江苏连云港人, 博士, 主要研究方向: 模糊建模、智能控制等, E-ma il:lzy69813@gmail.com。
更新日期/Last Update: 2007-08-30