| With the rapid development of social economy, the contradiction between human and water environment is more and more serious. The problem such as the shortage of water resources, water eutrophication has become more and more important to be focus on. According to the studies, the point source pollution is under the effective control, we want to make non-point pollution for effective prevention and governance. China’s ministry of environmental protection plans to develop a water environment system suitable for most waters, it can do the simulation and calculation of hydrological processes, sediment, nutrient load and so on.According to the condition of hydrology started late in China and the related monitoring data is not complete, we take the GWLF for the calculation of pollutions. But as a result of semi-empirical non-point load model of the localization of environment related parameters have certain requirements, its reliance on environmental parameters will cause large errors, we choose the genetic algorithm which is no special requirements about the number of parameters and the concave and convex of the function.Genetic algorithm process is calculated by using the ideas of biological evolution, through selection, crossover and mutation calibration of the three genetic operators, the choice of different genetic operator has great influence on the result. This article shows the experimental comparison of selection, crossover and mutation of the genetic operators to determine the parameters of genetic operators, in order to adapt to the hydrological processes, pollutants and nutrient load parameters calibration requirements, we use the adaptive crossover operator scheme design to deal with different initial conditions.Finally by using the test data of yuqiao reservoir in Tianjin to simulation, after the calibration of using adaptive genetic algorithm, its final decision coefficient can reach 0.91, it increased by 5% than the original by using experiential local data is 0.87. |
PDF Full Text Request