2007 4 4 :100026788 (2007) 0420114206, (, 430074) :,,,,,,GIS.,,. : ; ; ; ; : TP391 ;P338 : A Development of Combinatorial Intelligentized Decision2Making Support System and Its Utilization in Runoff Forecasting System HUANG Mu2tao, TIAN Yong ( The Center of Digital Engineering,Huazhong University of Science and Technology, Wuhan 430074,China) Abstract : In order to solve the current main problems of Decision2Making Support System (DSS), such as shortage of enough information resources support, difficulties in obtaining and expressing the expert knowledge, trouble of extracting useful information from mass isomeric data and so on, we have developed the combined intelligent information processor based on recombining and improving the traditional modeling methods including artificial neural network(ann),wavelet transformation(wt), genetic algoritnm( GA). In addition, an intelligent DDS is established by integrating the combined intelligent information processor with DSS, which can extract out useful knowledge and information from mass database management system automatically and build intelligent modeling system. Furthmore, based on the GIS platform, the integration of all module in the intelligent DDS is completed using component technology, and the integrated scheme is described clearly in this paper. The example of combinatorial intelligentized DSS is given for the water resources management DSS in a irrigation catchment, including main interfaces based on GIS and the simulation results of rainfall2runoff forecasting module. Key words : neural network ; genetic algorithm ; wavelet transformation ; intelligent decision ; component technology,,,: [1 ]., [2,3 ], GIS,,. 1 :, :2006201212 :(1973 - ),,,.
4 115,,. : p { ( x ( t), y ( t) ) x R m, y R n, t = 1,2,, p} (1),, x ( t), y ( t),,: F : R m R n.,m, n, u.,, Sigmoid. :,f Sigmoid,f ( x) = ^y k ( t) = f u v jk a, b j = 1 m w ij x i ( t) + j + r k i = 1, (2) 1 1 + e - x, k = 1,2,, n ; t = 1,2,, p, x i, ^y k, w ij i j, v jk j k, j j, r k k. net j = a j, b j. m w ij x i + j, i = 1 a, b ( net j ) = net j - b j, a j (a j, b j w ij v jk j, r k ).,,,.,,., E = 1 p n 2 t = 1 k = 1 [ y k ( t) - ^y k ( t) ] 2, (3),: p 1 n min E( w, v,, r) = 1 2 [ y k ( t) - ^y k ( t) ] 2 t = 1 k = 1, (4) s. t. w R mu, v R un,r u, r R n,,(3), E,. ( 1) : 1. 1),. 2). (),. 3),p ( m) ( n). 4),,,. 5). 6). MSE RMSE MAE. 2. 1),,. a, b ( x) Morlet,( x) = cos (1175 x) exp ( - x 2 Π2). 2). < 1 < 2 : < 1 = { ( x ( t), y ( t) ) x R m, y R n, t = 1,2,, p 1 ; p 1 p}, (5)
116 2007 4 < 2 = { ( x ( t), y ( t) ) x R m, y R n, t = p 1 + 1, p 1 + 2,, p}, (6) < 1, < 2 2.,E ;E <( ),,3 ; 3,. 1). N,, N. N. 2)., ()..,. w ij v jk j r k a j, b j,,.. : h 1 w h 11 w h m1 a h 1 b h 1 h uw h 1 u w h mu a h ub h u r o 1 v o 11 w h mu a h ub h u r o kv o un,, h, o. 3)..,. : F( w, v,, r) = 1Π(1 + E). (7) 4) (7). 5).,,., (,).. 6).,. 7) 4) 6),,, K( ).. 8) K,, 1 4.,.,,, 3,,.
4 117 2 211,, Π Π3, Oracle9i, ArcGIS 9 ArcSDE,, ArcIMS Web, ArcObject,. 212 (COMΠDCOM),, (DLL) ActiveX (OCX) ( EXE),,. COMΠDCOM,,,., DLL OCX EXE,.,.,,,,. 213 :, ;.,,, :, ;,, [4 ]. :,(DLL) (COM).,;.,,,,. 2. 2 214 Visual Studio. C #, 3,,,,GIS,. GIS,,. :,(DLL) (Oracle ),(DLL),,,, ;,., ;
118 2007 4,,,,,,,.,GIS ActiveX 2 3,. 3 3,,;. 4 Visual C #, Matlab7,COM. 7 1 20,.. BP Morlet,5,100 50 016 0101. 56 BP. 1,, BP. 1 BP BP ( %) ( %) 1d 0. 37 0. 364 1. 6 0. 371 0. 27 2d 0. 37 0. 361 2. 4 0. 366 1. 1 3d 0. 4 0. 379 5. 25 0. 394 1. 5 4d 21. 4 19. 98 6. 64 21. 02 1. 78 5d 2. 41 2. 2 9. 72 2. 39 0. 83 6d 0. 75 0. 67 8. 71 0. 73 2. 67 7d 0. 65 0. 71 6. 15 0. 64 1. 54 4
4 119,,,. 4 5 BP 6 : [1 ],,. [J ].,2005,27 (8) :30-31. Luo Fengman, Zhou Xinzhi, Gu Zhongbi. Structural design of the intelligent decision2making support system of water resource programming[j ]. Yellow River,2005,27 (8) :30-31. [2 ],,. [J ]. (),2005,26 (3) :118-120. Zeng Shuiling,Song Luanjiao,Xu Weihong. Struction optimization of neural network based on genetic algorithm[j ]. Jilin University Journal (Natural Science),2005,26 (3) :118-120. [3 ],,,. GIS [J ].,2002,33 (7) :27-30. Lu Hui,Tian Fuqiang,Hu Heping,et al. Irrigation decision2making support system based on genetic algorithm and GIS technique[j ]. Water Resources and Hydroelectric Engineering,2002,33 (7) :27-30. [4 ],. IDSS [J ].,2005,26 (8) :2083-2085. Gao Guangyao,Wang Shiqin. Research on the IDSS knowledge base construction based on the integration technique of neural network [J ]. Computer Engineering and Design,2005,26 (8) :2083-2085.