Projection pursuit grade evaluation model based on maximum information entropy principle for river health system

In order to rationally determine each index weight and effectively deal with fiazziness and randomness existing in evaluation process, projection pursuit grade evaluation model based on maximum information entropy principle, named PPE for short, is proposed as a new evaluation approach. Optimal projection direction of PPE using accelerating genetic algorithm reflects the objective weights of evaluation indexes, hereby the rationality of original evaluation criterion can be verified and the major indexes of influencing river health can be identified. The results of PPE applied to evaluating a river health in the South of China show that the calculated membership degree value of the river health fuzzy set is 0.6115, so the river belongs to the grade of “health” and approaches the grade of “inferior health”; there is only objective weight relatively small value of the index “river blanking probability” in the 25 evaluation indexes, and the objective weight value of other indexes approaches 0.04; the major indexes of influencing the river health are “the collecting rate of the fee of water resource”, “self-purification rate of water body”, “the index of regulating ability”, “change rate of sediment transport capacity”, “GDP of per cubic meter water”, “change rate of fish kind”, “perfect rate of nonstructural measures of flood control” and “sewage handling rate”, respectively; it is the important working content of the water security management of this river at present to improve the river health degrees of these indexes.