Optimal water allocation integrated with water supply, replenishment, and spill in the in-series reservoir based on an improved decomposition and dynamic programming aggregation method

Abstract

In humid regions with the monsoon climate, seasonal water shortages and water spills occur alternately because of uneven temporal and spatial distributions of water resources. An optimization model for the in-series reservoir (ISR) with replenishment pumping stations was developed to obtain the minimum annual sum of water shortage and systematically considered the reservoir operation rule of water spill and replenishment. This model features multiple dimensions; dynamic programming (DP) may cause a ‘curse of dimensions’, while the decomposition-coordination method has difficulty in judging logic conditions in the reservoir operation rules. So, an improved decomposition and DP aggregation (DDPA) method was proposed. The proposed model and the method were applied to a real case in the humid region of southern China. Compared to a conventional scheduling method, the water supply was increased by 0.8% and replenishment was reduced by 2.5%. Moreover, a comparison between DDPA and six heuristic algorithms was discussed. All heuristic algorithms' objective function values only obtained local optimal solutions, and the water shortage of the system was 0.12–20.5%. The obtained results demonstrated that DDPA was the better choice for highly complex multi-reservoir systems. The proposed optimization algorithm enriched the optimization theory of multi-dimensional and multi-variable complex systems. HIGHLIGHTS An optimization model integrated with water supply, replenishment, and spill in the in-series reservoir (ISR) was proposed.; An improved decomposition and dynamic programming aggregation (DDPA) method was proposed.; The performance of DDPA was superior to the six heuristic algorithms.; The water utilization efficiencies of the ISR were improved.;

Keywords

• dynamic programming
• heuristic algorithms
• in-series reservoir
• multi-dimensional
• optimal scheduling
• water resources allocation