Exploration For Water Quality Assessment And Prediction Based On Neural Networks And Artificial Bee Colony Algorithm

Water is an important resource for human survival and development, however, with thedevelopment of economy, water shortage and water pollution has become one of the fatalissues facing China’s economic and social development. Evaluation and prediction of waterquality is an important part of the study of water environment, and it’s an important means ofwater management and maintenance.The artificial bee colony algorithm is a swarm intelligence algorithm, it’s motivated byintelligent behaviour of honey bees. Artificial neural network is a data processing systemwhose architecture essentially mimics the biological system of a brain. The BP neural networkis one of the most widely used. But the BP neural network has some shortcomings, such asslow rate of convergence, sensitive to the initial value, easy to fall into local minima and so on.In the paper, artificial bee colony algorithm and BP neural network was combined, artificialbee colony algorithm was used to find the optimal weights and thresholds of BP neuralnetwork, weight and thresholds problem of BP neural network was transformed to the processof searching the best nectar for honey bees. A new method (ABC-BP) was proposed, and itwas used to the establishment of water quality evaluation and prediction models.In the paper, firstly, artificial neural networks and artificial bee colony algorithm wasdetailed. And then the BP neural network and artificial bee colony algorithm applied in waterquality evaluation and prediction was described, and finally realization of water qualityevaluation and prediction in Visual C++was briefly explanation.Water quality assessment is based on our country’s “surface water quality standards”, BPneural network and ABC-BP approach were used to build water quality assessment model, onten groups measured data of Weihe River in2000-2006year, the evaluation results werecompared and analysis, concluded that the water quality evaluation model established by theABC-BP method can get accurate and stable water quality evaluation results. Water qualityprediction took the prediction of dissolve oxygen for example, select ten water qualityobjectives which were relevant to dissolve oxygen, ABC-BP algorithm was used to establishwater quality prediction model. In order to study the feasibility and effective, it was comparedwith another algorithms—genetic algorithm optimization BP neural network. Through a largenumber of simulation results and analysis, it is obtained that ABC-BP algorithm can be usedto establish water quality prediction models, and has higher prediction accuracy, and also theprediction error is relatively stable.