Abstract: Regional water resources carrying capacity in Yunnan Province is assessed by integrating maximum entropy with projection pursuit (MEPP) premised on an index system and rating standard involving water resources, economy and society, as well as ecological system. Data in the recent decade from 2006 to 2015 is taken as a case study. A MEPP model is established with its target functions constructed by random samples between the thresholds of the target classification criteria and its optimal projection direction optimized by shuffled frog leaping algorithm (SFLA). Moreover, models of biogeographic optimization (BBO)-MEPP, Harmony Search (HS)-MEPP and Particle Swarm Optimization (PSO)-MEPP are built as well for comparison. Results demonstrate that the SFLA is of better ability of searching optimal values globally, with its optimum value, most inferior value, mean value and standard deviation superior to those of BBO, HS and PSO. According to the results of SFLA-MEPP model, the water resources carrying capacity in Yunnan Province from 2006 to 2007 and from 2011 to 2012 are overall loadable, and in the other years are rated as loadable. The carrying capacity of water resources in Yunnan Province had been enhancing with time in 2006-2015, yet the trend was insignificant. The results of SFLA-MEPP model are identical with those of BBO-MEPP model with different ranking orders; but are different with those of HS-MEPP model and PSO-MEPP model.

Key words: water resources carrying capacity, Yunnan Province, projection pursuit, maximum entropy, shuffled frog leaping algorithm