Deriving Water Allocation Schemes for Interbasin Water Transfer Projects Using a Novel Multiobjective Cuckoo Search Algorithm

Abstract

Interbasin water transfer projects (IBWTPs) have become an important measure to alleviate regional water stresses by artificially regulating water resources between water-rich and water-scarce areas. However, there are many conflicting objectives and complex practical operational constraints that complicate the derivation of water allocation schemes for IBWTPs. Here, we propose a multiobjective optimization methodology for deriving water allocation schemes for IBWTPs, which includes three modules: (1) formulating a multiobjective optimal water allocation problem for IBWTPs based on practical operational constraints; (2) proposing a novel multiobjective cuckoo search (NMOCS) algorithm and combining it with a simulation-optimization approach to solve the water allocation problem; and (3) filtering the Pareto solutions using the analytic hierarchy process (AHP)-entropy method. Based on the multiobjective cuckoo search (MOCS) algorithm, the NMOCS employs four improvement strategies, including a population initialization strategy, flock search strategy, multistrategy external archive maintenance strategy, and adaptive parameters, to improve the convergence property and diversity of solutions. To test the performance of the NMOCS, we considered the MOCS and four widely used multiobjective evolutionary algorithms (MOEAs) as a comparison and employed these six MOEAs to solve 11 multiobjective mathematical benchmark problems as well as the water allocation problems of the Jiangsu Province, China, section of the South-to-North Water Diversion Project under normal, dry, and extremely dry hydrological conditions. The results show that the NMOCS outperformed other MOEAs in handling multiobjective mathematical benchmark problems, especially in ZDT1, ZDT2, ZDT6, DTLZ2, DTLZ4, and DTLZ7. Compared with other MOEAs, the NMOCS did not always capture the highest percentage of the reference water allocation schemes, but it provided more than 18% (greater than one-sixth) of effective water allocation schemes for all hydrologic conditions. Meanwhile, compared with optimal water allocation schemes derived from other MOEAs, the NMOCS effectively improved the operational performance under normal and drought hydrological conditions, especially in the total water pumping and the water withdrawn from the Yangtze River. This research can help to update our understanding of MOEAs, particularly the MOCS, and serve as a reference for better-allocating water resources in IBWTPs.