Development of a web-based software for realizing the optimal water quality monitoring network for a river system using genetic algorithm and artificial neural network

Despite decades of activities in the field of water quality, the locations of water quality stations are still in many cases empirically determined. The main purpose of this study is to develop a web-based application for the effective design of a network of water quality monitoring stations in rivers using the NSGA-II, geographic information system, and open-source tools. As a result, users can easily use this program with only Internet and web browser access without the complexity of using desktop-based applications. For this purpose, in the design of the river water quality monitoring stations network, the general objectives of the monitoring were determined and the fitness function was designed. Then, a web-based application was implemented using the C# programming language in the Visual Studio environment. This web-based application receives the necessary data in Shapefile format and displays the optimal locations of water monitoring stations on the web and on the map after applying NSGA-II. To find the optimal locations of the water quality monitoring stations, the NSGA-II was run for different generations. Pareto optimal solutions were obtained in the 2500 generation. With the results of NSGA-II being compared with the results of Analytical Hierarchical Process and Fuzzy Logic methods, it was observed that the results of these methods were largely compatible. In the next step, a GA optimized multilayer Perceptron neural network (MLP) was used to accelerate the finding of optimal locations for river water quality monitoring stations, and the MLP results were evaluated by statistical indicators with the NSGA-II results. MLP results with momentum learning rule and Conjugate Gradient learning rule showed similar results to NSGA-II results with errors of 0.185 and 0.163.