Optimal river basin water resources allocation considering multiple water sources joint scheduling: A bi-level multi-objective programming with copula-based interval-bistochastic information

The uncertainty of hydrological variables and socio-economic parameters presents new challenges for river basin water resources allocation (RBWRA). Motivated by the pressing need for more efficient and eco-friendly RBWRA solutions, a bi-level multi-objective interval-bistochastic programming (BLMOIBSP) model is proposed. The model aims to achieve optimal balance among efficiency, eco-friendliness and equity, collectively called “3E” in this paper. The basin authority, serving as the leader in RBWRA, prioritizes efficiency and eco-friendliness, aiming to maximize economic benefit and minimize total environmental pollutants. Conversely, the sub-areas, as followers, focus more on benefits, striving to maximize total benefits. Taking into account the needs of each stakeholder in the bi-level structure actually facilitates equity goals in RBWRA. Additionally, based on the joint scheduling of multiple water sources, the model considers dual stochastic hydrological variables and multiple uncertain parameters. The copula function captures the dependence between hydrological stochastic variables. Meanwhile, water demand is set as an interval parameter to address uncertainty, handled by an interval two-stage stochastic programming (ITSP) method. Fusing with a chaotic Aquila Optimizer (CAO) algorithm, a bi-level interactive approach based on satisfactory degree (SD) is employed to solve the model and ensure equity in RBWRA. Furthermore, the case of the Dongjiang River Basin in Guangdong Province, China, is presented in conjunction with a multi-scenario analysis to validate the practicality and effectiveness of the proposed model and its associated problem-solving methods. The results indicate that the proposed model effectively ensures the “3E” objectives of RBWRA. Besides that, ITSP has been demonstrated to reduce the risks associated with the uncertainty in actual water allocation in RBWRA. The proposed bi-level model balances the levels better than a single-level model. When solving the model, the improved CAO algorithm demonstrates better convergence and stability. Finally, the strengths of the proposed model and the potential challenges in practical scenarios are remarked and future research directions are indicated.