Abstract: Based on the associative memory of discrete Hopfield neural network, a model to  comprehensively assess the eutrophication level of lakes and reservoirs is established. Twenty-four lakes and reservoirs in China are evaluated through this model, and the results are compared with those of  projection pursuit method,  score index method, and LM-BP network method. The results show that discrete Hopfield neural network is simple, intuitive, and easy to implement, with only a few  iterations leading to satisfactory and objective results. However, not all eutrophication level assessments could be achieved through general discrete Hopfield neural network. When there is a big difference between each single index (factor), correct assessment could not be achieved.

Key words: eutrophication assessment, ANN (artificial neural network), Hopfield network, lakes and reservoirs