Multi-objective optimization in real-time operation of rainwater harvesting systems

Increased population growth and urbanization have brought critical challenges to urban water systems, including water scarcity and environmental degradation. To address the problems, real-time controlled rainwater storages are now being used to reduce flooding by intercepting rainfall, while also providing an alternate water supply and actively restoring baseflow to improve biodiversity outcomes. These benefits can be enhanced when the storages are managed as an optimized network. This paper proposes a multi-objective-optimization-based strategy utilizing mixed integer linear programming and compromise programming to control a network of rainwater storages. The proposed strategy is observed to substantially reduce storage overflow, improve stream baseflow, and fulfill most of the domestic non-potable water demand. It shows a clear advantage over the NSGA II-based strategy, indicating the effectiveness of mathematical programming with scalarization techniques in solving multi-objective problems.