Stochastic multi-attribute group decision-making for long-term comprehensive operation of cascade reservoirs under multiple uncertainties

Abstract

Study region: The Three Gorges cascade (TGC) reservoirs, China. Study focus: Since uncertainties inherent in inflow forecasting, elevation-storage, discharge capacity, flood routing, and the preferences of decision makers (DMs) persist in the attribute measurements (AMs) and weights (AWs), problems where multiple DMs select the best compromise scheme for the long-term comprehensive operation of cascade reservoirs considering numerous attributes become a stochastic multi-attribute group decision-making (MAGDM) process. To represent uncertain AMs and AWs and consider the bounded rationality of DMs, we propose a stochastic MAGDM with integrated stochastic multicriteria acceptability analysis (SMAA) and gray cumulative prospect theory (GCPT), namely, SMAA-GCPT. After quantifying the uncertainties, SMAA-GCPT is used to model uncertain AMs and imperfect AWs with calculable probability distributions and map the sampled AMs and AWs to utility values to reflect the overall performance assuming bounded rationality. New hydrological insights for the region: Experiments on the TGC indicate that SMAA-GCPT can effectively differentiate alternative schemes and select the best compromise scheme to harmonize the conflicting benefits and risks of the TGC, with higher reliability than SMAA-2 and SMAA-GRA. The influence of uncertain AMs on the reliability of SMAA-GCPT is more obvious than that of uncertain AWs. The decision-making error risk is primarily sensitive to the inflow forecasting uncertainty, followed by the flood routing uncertainty, and is less affected by the uncertainties in elevation-storage and discharge capacity curves.

Keywords

• Stochastic multi-attribute group decision-making
• Uncertainty
• Multi-objective optimization
• Combination weighting method
• Gray cumulative prospect theory