Spatial quality evaluation for drinking water based on GIS and ant colony clustering algorithm

作     者：侯景伟 米文宝 李陇堂 

作者机构：School of Resource and EnvironmentNingxia University 

出 版 物：《Journal of Central South University》 (中南大学学报（英文版）)

年 卷 期：2014年第21卷第3期

页      面：1051-1057页

核心收录：

学科分类：12[管理学] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 081104[工学-模式识别与智能系统] 08[工学] 0835[工学-软件工程] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：Projects(41161020,41261026) supported by the National Natural Science Foundation of China Project(BQD2012013) supported by the Research starting Funds for Imported Talents,Ningxia University,China Project(ZR1209) supported by the Natural Science Funds,Ningxia University,China Project(NGY2013005) supported by the Key Science Project of Colleges and Universities in Ningxia,China 

主　　题：geographical information system (GIS) ant colony clustering algorithm (ACCA) quality evaluation drinking water spatial analysis 

摘      要：To develop a better approach for spatial evaluation of drinking water quality, an intelligent evaluation method integrating a geographical information system(GIS) and an ant colony clustering algorithm(ACCA) was used. Drinking water samples from 29 wells in Zhenping County, China, were collected and analyzed. 35 parameters on water quality were selected, such as chloride concentration, sulphate concentration, total hardness, nitrate concentration, fluoride concentration, turbidity, pH, chromium concentration, COD, bacterium amount, total coliforms and color. The best spatial interpolation methods for the 35 parameters were found and selected from all types of interpolation methods in GIS environment according to the minimum cross-validation errors. The ACCA was improved through three strategies, namely mixed distance function, average similitude degree and probability conversion functions. Then, the ACCA was carried out to obtain different water quality grades in the GIS environment. In the end, the result from the ACCA was compared with those from the competitive Hopfield neural network(CHNN) to validate the feasibility and effectiveness of the ACCA according to three evaluation indexes, which are stochastic sampling method, pixel amount and convergence speed. It is shown that the spatial water quality grades obtained from the ACCA were more effective, accurate and intelligent than those obtained from the CHNN.