PSO优化的模糊神经网络在管道泄漏检测中的应用研究
详细信息    本馆镜像全文|  推荐本文 |  |   获取CNKI官网全文
摘要
随着石油管道运输业的不断发展,管道在国民经济中的地位越来越重要,而管道泄漏事故也时有发生。管道泄漏不仅会造成能源的浪费,带来巨大的经济损失,而且会污染环境,甚至会威胁到人民生命财产的安全。因此,长输管道的安全运行受到了越来越多的关注,管道泄漏检测定位成为当前重要的研究课题。管道泄漏检测与定位方法诸多,其中基于知识的方法由于无需为复杂的管道系统建模、诊断结果受不确定因素影响较小、且具有定性和定量分析的双重功效,近年来受到业界的广泛关注。基于此,本文以某实际管道运行的数据为背景,在前期基于模糊BP神经网络的检漏方法研究的基础上,针对存在的一些问题,结合近几年一些新技术新成果,对管道泄漏的检测与估计方法做了进一步的研究。其主要内容如下:
     1.针对模糊BP神经网络存在收敛速度慢、易陷入局部最小等问题,本文构建了基于PSO优化了的模糊RBF神经网络的管道泄漏检测方法,通过现场真实数据的仿真研究,显示出模糊RBF网络较模糊BP网络对管道泄漏检测具有更高的可靠性和准确性。
     2.针对一般模糊神经网络运算函数对模糊逻辑融合不足,且权值的优化易陷入局部最优等问题,文中将一种以广义概率积、广义概率和模糊算子替代其运算函数的模糊神经网络用于管道的泄漏检测,并用发散-收缩PSO(DCPSO)算法优化了该模糊神经网络的权值,建立了DCPSO-FNN神经网络;并在固定权值的基础上,分别采用基本PSO和DCPSO优化广义模糊算子中悲观度、乐观度参数,经仿真研究,该方法具有更高的检测与估计精度。
     3.考虑到网络的结构比较庞大,并且部分权值的影响非常微弱,为提高网络的检测效率,本文依据一定的裁剪算法,将该DCPSO-FNN神经网络进行了裁剪,得到了更为简捷的模糊神经网络,并以现场数据仿真验证了裁剪后的网络可以较为精准地对泄漏进行检测与估计,从而为实际应用提供了便利。
'With the development of pipeline transportation, pipeline is getting more and more important in national economy status,but the pipeline leakage accident also sometimes occurs.The pipeline leakage will not only cause the wastage of power resource and the huge economic losses, but also pollute the environment, threat the safety of people.Therefore, long pipeline's safe operation receives more and more attention, the pipeline leakage detection and localization have become the current important research subject. There are many methods in pipeline leakage detection and localization. The method based on knowledge does not need to establish the complex model of pipeline system, and its diagnosis result is not affected by the uncertainty factors,and it has the qualitative and the quantitative analysis dual effects, so it receives widespread attention in the field in recent years. Based on this, the methods of pipeline leakage detection and estimation have been further studied, under the background of the actual data in the pipeline leakage network, in the earlier period based on the fuzzy BP neural network's leak detection method study's foundation, in view of existence some questions,with the union of some new technologies and new achievements in recent years,in this work.The main contents as the following aspects:
     1.In view of the fuzzy BP neural network exists convergence rate slow and easy to fall into partially smallest and so on questions,this article has constructed the fuzzy RBF neural network pipeline leakage detection method optimized by PSO, the real data's simulation research confirmed that the fuzzy RBF neural network has a better reliability and the accuracy than the fuzzy BP network.
     2.In view of general fuzzy neural network operating function to fuzzy logic fusion insufficiency, and the weight optimization easy to fall into partially most excellent and so on. In the article, the fuzzy neural network that the fuzzy operator of the general probability, the probability and generalized substitutes its operating function is used in pipeline leakage detection, using the divergent-convergent PSO(DCPSO) algorithm to optimize the weights of this fuzzy neural network, established the DCPSO-FNN neural network; based on the fixed weights, using basic PSO and DCPSO separately to optimize the pessimism, happy parameter of the generalized fuzzy operator, the simulation result indicated that this method has the higher detection and estimation precision.
     3.Considered the structure of the network is quite huge, and the partial weights' influence are vary weak, in order to enhance the detection efficiency of the network, this article has carried on cutting out this DCPSO-FNN neural network based on certain cutting out algorithm,obtained the more simpler and direct fuzzy neural network.After the real data simulation, confirmed that the cutting out network has the accuracy and the reliability to pipeline leakage detection and estimation, thus has provided the convenience for the practical application.
引文
[1]王霖.我国管道运输已成为五大运输行业之一[J].石油消息,1995,20(1):1-5.
    [2]潘家华.管线安全的科学管理[J].管道运输,1997,2(3):1-9.
    [3]别沁,郑云萍等.国内外油气输送管道泄漏检测技术及发展趋势[J].石油工程建设,2007,33(3):19-23.
    [4]杨军.油(气)管道泄漏检测方法的研究[M].北京:清华大学,1994.
    [5]周东华,叶银忠.现代故障诊断与容错控制[M].北京:清华大学出版社,6-10.
    [6]陈春刚,王毅,杨振坤.长输油气管道泄漏检测技术综述[J].石油与天然气化工,2002,31(1):52-54.
    [7]高宏扬.管道泄漏检测技术的应用与发展[J].油气储运,2004,23(11):1-2.
    [8]徐洁,丁金婷,江皓.管道泄漏检测方法综述[J].管道技术与设备,2004,4:14-16.
    [9]李炜,朱芸.长输管线泄漏检测与定位方法分析[J].天然气工业,2005,25(6):105-109.
    [10]瞿曌.液体输送管道泄漏的检测与定位[J].油气储运,2005,24(4):14-18.
    [11]王雪亮,苏欣.油气管道泄漏检测技术综述[J].天然气与石油,2007.6,25(3):19-23.
    [12]唐恂,张琳等.长输管道泄漏检测技术发展现状[J].油气储运,2007,26(7)11-14,29.
    [13]王延年,朱强,赵则祥.长输油气管道泄漏检测方法研究进展[J].石油机械,2007,35(5):49-53.
    [14]王效东,黄坤,朱小华,等.油气管道泄漏检测技术发展现状[J].管道技术与设备,2008,24-26.
    [15]Wuu-Wen Lin.The Feasibility Study of A Leak Detection System for An Underwater Pipeline[J].Institute of Technology,2003,4:2100-2104.
    [16]王潜龙,等.基于声发射与小波包理论的压力管道泄漏检测[J].西安交通大学学报,2003,37(5):515-518.
    [17]龚斌,包日东,闻邦椿等.压力管道泄漏点的新型声发射定位研究[J].化工机械,2005,32(5):291-293.
    [18]冯健.流体输送管道泄漏智能诊断与定位方法的研究[D].沈阳:东北大学博士学位论文,2005.
    [19]袁朝庆,刘燕,才英俊等.利用光纤温度传感系统检测天然气管道泄漏[J].天然气工业,2006,26(8),117-119.
    [20]王朝辉,张来斌等.声发射技术在管道泄漏检测中的应用[J].中国石油大学学报(自然科学版),2007年10月,31(5):87-90.
    [21]李兆南,龚斌,林木,殷天舟.压力管道泄漏声发射信号频谱特性实验研究[J].声学技术,2007.6,26(3):422-426.
    [22]张立,侯迪波,张光新,周泽魁.基于多声学传感器融合的管道泄漏检测方法研究[J].传感技术学报,2007.5,20(5):1176-1179.
    [23]李善春,郭福平,王为松.压力管道泄漏声发射监测试验研究[J].无损检测,2007,29(2):74-76,79.
    [24]张淑清,李二伟等等.DOFRPTS及小波分析在管道泄漏检测定位中的应用[C].第五届全国信息技术获取与处理学术会议,2007年.
    [25]Shih-Chu Huang, Wuu-Wen Lin.et.1.Fiber optic in-line distributed sensor for detection and localization of the pipeline leaks[J]. Sensors and Actuators A 135(2007)570-579.
    [26]杭利军,何存富,吴斌.基于Sagnac光纤干涉仪的管道泄漏检测和定位技术[J].光学技术,2007.9,33(5):651-653.
    [27]杭利军,何存富,吴斌.一种新的直线型Sagnac光纤干涉仪管道泄漏检测系统及其定位技术[J].中国激光,2007.6,34(6):820-824.
    [28]陈志刚,张来斌,王朝晖,梁伟.基于分布式光纤传感器的输气管道泄漏检测方法[J].传感器与微系统,2007,2(7):108-110.
    [29]袁朝庆,刘燕,才英俊,靳志强.利用光纤温度传感系统检测天然气管道泄漏[J].天然气工业,2006,26(8):117-119.
    [30]周琰,靳世久,张昀超,孙立瑛.分布式光纤管道泄漏检测和定位技术[J].石油学报,2006.3,27(2):121-124.
    [31]陈华立,叶昊.基于图像处理的管道泄漏检测与定位[J].清华大学学报,2005,45(1):119-122.
    [32]Toshio Fukuda.Pipeline inspection and maintenance by application of computer data processing and robotic technology[J].Computers in industry,1986,12(1):5-23.
    [33]John Pitchford.In-line inspection-is sufficient data being collected, int[J].Journal of Pipes & Pipelines,1995,40(6-8):28-30.
    [34]偶国富,朱祖超等.埋地金属管道地面电磁检测技术研究[J].仪器仪表学报,2007.2,28(2):258-263.
    [35]张连文.管道泄漏检测技术及评价[J].油气田地面工程,2003,22(4):1-3.
    [36]靳世久,姜弘彦.强环境噪声下地下管道泄漏检测[J].天津大学学报,1994,27(6):782-787.
    [37]彭柯等.基于SCADA系统的泄漏检测与定位[J].化工自动化及仪表,2004,31(1):50-51.
    [38]汪冈伟,宋建河等.利用SCADA系统实现管道泄漏检测与定位[J].油气储运,2004,23(10):46-50.
    [39]Dr. R.S. Whaley et al.Tutorial on Software Based Leak Detection Techniques[J].Scientific Software-Intercomp.Pipeline Simulation Interest Group, Oct 1992:1-19.
    [40]Jim. C. P. Liou. Leak detection by mass effective for Norman wells line [J].Oil & Gas Journal,1996(3):69-74.
    [41]American Petroleum Institute(API).Computational Pipeline Monitoring[J],New York:API Publication 1130,1995.
    [42]T.Kiuchi.A Leak Localization Method of Pipeline by Means of Fluid Transient Model[J].Journal of Energy Resources,1993,9:162-167.
    [43]陈春刚.基于动态质量平衡的管道泄漏检测方法研究[D].西安:西安交通大学硕士学位论文,2003.
    [44]胡坚译.用质量平衡方法进行管道检漏[J].油气储运,1997,16(12):55-58.
    [45]蒋仕章,蒲家宁.用动态质量平衡原理进行管道检漏的精度分析[J].油气储运,2000,19(2):12-13.
    [46]程家铭,张汉国.输油管道负压波法测漏原理及实现[J].石油机械,2002,30(9):28-30.
    [47]邓鸿英,杨振坤等.负压波管道泄漏检测与定位技术[J].油气储运,2003,22(7):30-33.
    [48]宋立军,杨拥民,熊先锋.基于负压波分析的油管检漏技术研究[J].测控技术,2003,22(2):44-46.
    [49]冯健,张化光.管道泄漏计算机在线检测系统及其算法实现[J].控制与决策,2004.4,19(4):377-382.
    [50]刘昕宇.负压波用于管道泄漏检测的研究[D].北京:中国石油大学硕士学位论文,2005.
    [51]臧国军,张鸣远,罗毓珊,苏欣,张琳.靖—咸原油管线泄漏报警检测系统试验研究[J].西南石油大学学报,2007.8,29(4):161-163.
    [52]廉小亲,苏维均,田黎明.基于负压波法的输油管道泄漏检测定位系统[J].计算机工程与设计,2007.5,28(9):2199-2202.
    [53]王帮峰,陈仁文.基于应力波检测的输油管道泄漏定位监测系统[J].仪器仪表学报,2007.6,28(6):1012-1017.
    [54]Y. Gao, M. J. Brennan. A model of the correlation function of leak noise in buried plastic pipes.Journal of Sound and Vibration,2004, 277(2):133-148.
    [55]S B Beck.Pipeline System Identification through Cross-correlation Analysis.Department of Mechanical Engineering,2002,216:133-142.
    [56]王福明,胡志新.相关分析在油气输送管道检漏中的应用[J].油气田地面工程,1998.9,17(5):13,14,20.
    [57]梁伟,张来斌,王朝晖.信息缺失条件下管道泄漏信号识别研究[J].石油大学学报(自然科学版),2004,28(5):74-77,81.
    [58]王爱民,相关分析在管道泄漏技术中的应用[J].合肥工业大学学报(自然科学版),2004.9,27(9):1063-1065.
    [59]韩建,牟海维,王永涛,姜晓岚:相关分析法在输油管道泄漏检测和定位中的应用研究[J].核电子学与探测技术,2007.1,27(1):154-156.
    [60]张建利,佟凯,马放.相关分析法管道漏点定位系统的试验研究[J].哈尔滨工业大学学报,2007.6,39(6):875-878.
    [61]孙立瑛,李一博,曲志刚,靳世久,周琰.EMD信号分析方法的声发射管道泄漏检测研究[J].振动与冲击,2007,26(10):161-164.
    [62]余永辉,漆巧玲,彭宇兴.输油管道泄漏检测与定位系统中的相关分析法研究[J].电子技术应用,2009,(2):128-130.
    [63]王海生,叶昊,王桂增.基于小波分析的输油管道泄漏检测[J].信息与控制,2002.10,31(5):456-460.
    [64]周诗岽,吴志敏,赵玲.输油管道泄漏检测实验研究[J].辽宁石油化工大学学报,2004.12,24(4):39-42,50.
    [65]张东领,王树青,李华军.基于压力波的海底管道泄漏定位技术的研究[J].中国海洋大学学报,2006.12,第36卷增刊Ⅱ:189-192.
    [66]韩建,王永涛,牟海维,姜晓岚.小波变换在输油管道泄漏检测和定位中的应用[J].核电子学与探测技术,2006.3,26(2):212-214.
    [67]李帆,张丽娟.基于小波分析消噪技术的燃气管道泄漏检测与定位[J].流体机械,2006,34(12):47-51.
    [68]秦先勇,张来斌.基于OPC技术的管道泄漏诊断方法研究[J].石油机械,2006,34(7):46-48.
    [69]张东领,王树青,张敏.热输油管道泄漏定位技术研究[J].石油学报,2007.1,28(1):131-133,138.
    [70]潘霞,范世东,钟骏杰,胡琼.管道泄漏检测实验系统设计与实现[J].船海工程,2007.4,36(2):43-47.
    [71]白田卫,兰翼,杨自栋.小波变换法在输油管道泄漏检测中的应用研究[J].农业装备与车辆工程,2007年第8期:24-27,30.
    [72]张文奎,陈小惠.基于小波降噪的供水管道泄漏检测算法研究[J].电子测试,2009.10,第10期:6-9.
    [73]胡瑾秋,张来斌,梁伟,王朝晖.基于谐波小波分析的管道小泄漏诊断方法[J].中国石油大学学报(自然科学版),2009.8,33(4):118-124.
    [74]Zhang, X.J. Statistical leak detection in gas and liquid pipelines[J]. Pipes & Pipelines International,1993,38(4):26-29.
    [75]廖锐全,汪崎生,张柏年.井筒多相管流压力梯度计算新方法[J].江汉石油学院学报,1998.3,20(1):59-63.
    [76]侯秀林,邓宏文,谷丽冰.一个新的描述低渗透油藏油水两相渗流启动压力梯度的公式[J].石油天然气学报(江汉石油学院学报),2009,31(5):342-344.
    [77]庄朝文,富立,范跃祖.一种基于假设检验的多重渐消卡尔曼滤波[J].北京航空航天大学学报,2004.1,3(1):18-22.
    [78]白莉,岳前进,李洪升.基于水力瞬变与扩展卡尔曼滤波的管道流体监测与泄漏定位[J].计算力学学报,2005.12,22(6):739-743.
    [79]汪玮,邹涛,李少远.基于Unscented卡尔曼滤波器的管道泄漏的快速检测与定位[J].化工自动化及仪表,2008,35(1):40-44.
    [80]胡瑾秋,张来斌,王朝晖,梁伟.动态聚类算法在管道泄漏检测中的应用[J].石油机械,2007,35(2):31-35.
    [81]杨进,文玉梅,李平.基于盲系统辨识的供水管网泄漏定位系统研究[J].仪器仪表学报,2007年第28卷第8期:1456-1463.
    [82]Jin Yang, Yumei Wen, Ping Li.Leak location using blind system identification in water distribution pipelines[J].Journal of Sound and Vibration(2007):1-15.
    [83]刘恩斌,彭喜碧,李长俊等.基于瞬态模型的油气管道泄漏检测[J].天然气工业,2005(6):102-104.
    [84]吴小庆,王宇,江茂泽.关于输气管道的一个泄漏点的检测问题[J].西南石油学院学报,2006,28(3):27-29.
    [85]娄身强等.基于模型的管道泄漏监测抗扰动方法研究[J].控制工程,2007.9,14(5):551-554.
    [86]A. A. Carvalho, J M A Rebello. MFL Signals and Artificial Neural Networks Applied to Detection and Classification of Pipe Weld Defects[J].NDT & E International,2006,39(8):661-667.
    [87]Dongling Xu,Jun Liu.Inference and learning methodology of belief-rule-based expert system for pipeline leak detection[J]. Expert Systems with Applications.2007,32(1):103-113.
    [88]唐秀家,颜大椿.基于神经网络的管道泄漏检测方法及仪器[J].北京大学学报(自然科学版),1997.5,33(3):319-327.
    [89]伦淑娴,等.自适应模糊神经网络系统在管道泄漏检测中的应用[J].石油学报,2004,25(4):101-104.
    [90]孙晓松,谭兴强.基于小波分解的管道泄漏神经网络检测与定位[J].石油机械,2006,34(8):55-58.
    [91]王立坤,赵晋云,付松,谭东杰,李健,靳世久.基于神经网络的管道泄漏声波信号特征识别[J].仪器仪表学报,2006.6,27(6)增刊:2247-2249.
    [92]YAO Zhi-ying, PENG Guang-zheng, ZHANG Ya-li.Research on Gas Pipeline Leakage Detection and Localization Based on Diagonal Recurrent Neural Network[J].ACTA SCIENTIARUM NATURAL IUM UN IVERSITATIS SUNYATSEN I,2007.6,Vol.46,Sup.:93-94.
    [93]李炜,邝鹏,李阳.基于模糊神经网络的管道泄漏检测方法研究[J].计算机仿真,2009.2,26(2):190-192,232.
    [94]李炜,邝鹏.基于遗传优化的模糊神经网络在管道泄漏检测中的应用研究[J].科学技术与工程,2008.7,8(13):3490-3494.
    [95]崔谦,靳世久,李一博.模糊聚类分析方法在管道泄漏检测系统中的应用研究[J].电子测量与仪器学报,2006,20(2):60-62.
    [96]罗菁,倪建云.石油管道泄漏检测模糊识别研究[J].微计算机信息,2007年第23卷第12-2期:121-122.
    [97]郭新蕾,杨开林.基于瞬变流和遗传算法的管道泄漏辨识[J].计算力学学报,2009.10,26(5):664-669.
    [98]李俊花,孙昭晨,崔莉.一种新的长输管道泄漏监测方法[J].工程力学,2009.8,26(8):205-209,215.
    [99]Dong Yuhua, Yu Datao.Estimation of failure probability of oil and gas transmission pipelines by fuzzy fault tree analysis[J].Loss Prevention in the Process Industries,Vol.18,No.2,2005:83-88.
    [100]Jian Feng,Huaguang Zhang.Applications of fuzzy decision-making in pipeline leak localization[J].Fuzzy Systems,Vol.2,2004:599-603.
    [101]WAI-Rafei,RJBarnes.Underlying the performance of real-time software-based pipeline leak-detection systems[J].Pipes & Pipelines International,1999,44(6):44-51.
    [102]陈春刚,王毅等.利用假设检验理论进行管道泄漏检测[J].石油化工设备,2004,33(2):46-49.
    [103]崔谦等.基于序贯检验的管道泄漏检测方法[J].石油学报,2005,26(4):123-126.
    [104]李炜,苗丽,张耿.Boot strap与模糊聚类在管道泄漏检测中研究[J].微计算机信息(测控自动化),2007年第23卷第11-1期:292-294.
    [105]李炜,李阳,毛海杰.多序贯概率比检验在管道泄漏检测中的研究弹箭与制导学报,2007,27(4):266-269.
    [106]李炜,李阳.优化M-SPRT在管道泄漏检测与估计中的研究与实现[J].天然气工业,2009,(4):93-95.
    [107]Dr Jun Zhang,Micheal Twomey.Statistical pipeline leak detection techniques for all operating conditions[J].International Journal of Pressure Vessels and Piping,2001,229(2):42-45.
    [108]陈华立,叶昊.基于图像处理的管道泄漏检测与定位[J].清华大学学报(自然科学版),2005,45(1):119-122.
    [109]李俊花,孙昭晨,崔莉,贾旭.基于新息理论的变点泄漏监测检测及其在长输管道中的应用[J].自动化学报,2006.5,32(3):462-469.
    [110]方瑞明.基于聚类支持向量机的气体泄漏检测[J].仪器仪表学报,2007.11,28(11):2028-2033.
    [111]杨红英,叶昊等.关联维分析方法在管道泄漏定位中的应用[J].油气储运,2007,26(7):6-10.
    [112]文静,文玉梅等.地下管道泄漏检测定位中的时延估计方法[J].仪器仪表学 报,2007.7,28(7):1274-1280.
    [113]肖力,唐洪,徐刚.音波管道泄漏监测系统在新疆油田管道保护中的应用[J].计算机与应用化学,2009.9.28,26(9):1145-1147.
    [114]伍清,李保国等.管道泄漏实时监测系统的原理及其应用[J].油气储运,2003,22(8):38-40,44.
    [115]李炜,陈希平,毛海杰,潘纬.天然气管道泄漏点的定位检测方法研究[J].甘肃工业大学学报,2003.12,29(4):84-87.
    [116]李炜,朱芸,毛海杰.基于综合方法应用的管道泄漏检测研究[J].化工自动化及仪表,2005,32(2):51-53.
    [117]陈峰,付兴武.输油管道泄漏实时检测与定位的研究及应用[J].控制工程,2005.5,第12卷增刊:203-205.
    [118]路炜,文玉梅.供水管道泄漏定位中基于互谱的时延估计[J].仪器仪表学报,2007年3月,28(3):504-509.
    [119]Marco Ferrante, Bruno Brunone. Pipe system diagnosis and leak detection by unsteady-state tests.1.Wavelet analysis[J].Advances Water Resources 26(2003):107-116.
    [120]Henrique V.da Silva,Celso K.Morooka. Leak detection in petroleum pipelines using a fuzzy system[J].Journal of Petroleum Science and Engineering 49(2005):223-238.
    [121]陈志勇,李书臣等等.改进BP神经网络在管道泄漏检测中的应用[J].河南科技大学学报(自然科学版),2008.6,29(3):44-47.
    [122]王永涛,牟海维,韩建.基于小波包与神经网络的长输油管道泄漏检测方法[J].大庆石油学院学报,2006,30(4):129-131..
    [123]C. Verde. Multi-leak detection and isolation in fluid pipelines[J]. Control Engineering Practice,2001,9(6):673-682.
    [124]Cristina Verde, Nancy Visairo.Two leaks isolation in a pipeline by transient response[J].Advances in Water Resources,0(2007):711-1721.
    [125]杭利军,何存富,吴斌,蔡栋生.基于LabVIWE平台的锁相放大器在管道泄漏检测中的应用[J].北京工业大学学报,2007,34(3):241-244,264.
    [126]方伟,孙俊,须文波.一种多样性控制的粒子群优化算法[J].控制与决策,2008,23(8):863-868.
    [127]杨维,李歧强.粒子群优化算法综述[J].中国工程科学,2004,(6):87-94.
    [128]李宁,付国江,库少平,陈明俊.粒子群优化算法的发展与展望[J].武汉理工大学学报(信息与管理工程版),2005.4,27(2):26-29.
    [129]倪庆剑,邢汉承,张志政等.粒子群优化算法研究进展[J].模式识别与人工智能,2007.6,20(3):349-357.
    [130]Tang T H,Lin X, Li J R,Chen B S.A new fuzzy neural network approach for intelligent monitoring system[J].IFAC Transportation System, China,Greece,1997,691-696.
    [131]Riget J,Vesterstrainm J S. A diversity-guided particle swarm optimizer -The ARPSO[R].Denmark:University of Aarhus,2002.
    [132]Rusmas K Ursem. Diversity-guided evolutionary algorithms[C].Proc of Parallel Problem Solving from Nature VII(PPSN-2002).Heidelberg: Spring-Verlag,2002:462-471.