Rainfall-runoff simulation using improved evolution algorithm based on pattern search

Rainfall-runoff simulation is very important in practical environmental planning. To improve the computational accuracy for rainfall-runoff simulation, an improved evolution algorithm based on pattern search (IEAPS) is proposed, in which initial population are generated by chaos mapping and searching range is automatically renewed with the excellent individuals by pattern search. Its efficiency is verified by application of rainfall-runoff simulation for five rainfall events. Compared with standard binary-encoded genetic algorithm (SGA), pattern search algorithm (PSA), IEAPS has higher precision and rapider convergent speed. It is good for the global optimization in the practical rainfall-runoff simulation.