Optimising Urban Flood Resilience

TL;DR

This paper introduces a novel multi-objective optimisation tool combining hydrodynamic modelling and evolutionary algorithms to enhance urban flood resilience through optimal Blue-Green Infrastructure implementation.

Contribution

It presents a new optimisation framework that accurately evaluates flood mitigation solutions using dynamic hydrodynamic models and efficient algorithms, improving decision-making in urban flood management.

Findings

The tool accurately evaluates BGI effectiveness with dynamic hydrodynamic models.

It reduces computational load by minimising the number of simulations.

Validated convergence measures and benchmark comparisons demonstrate robustness.

Abstract

Due to the increasing frequency and severity of storm events, driven by the escalation of anthropogenic climate change and urban expansion, there is a requirement for increasingly efficient flood risk management strategies. While Blue-Green Infrastructure (BGI) offers a sustainable solution for managing flood risk, optimal implementation is challenging. To help overcome this challenge, this study presents a novel multi-objective optimisation tool that couples a state-of-the-art hydrodynamic model with a bespoke evolutionary algorithm. The use of a fully dynamic hydrodynamic model enables the tool to accurately evaluate the effectiveness of proposed BGI features with respect to property scale flood vulnerability and hazard analysis. This contrasts with alternative approaches which utilise simplified models, which can only reliably predict inundation extents, thus the proposed optimisation tool provides greater certainty regarding the optimality of the solutions. As a hydrodynamic simulation is required to evaluate each candidate solution, the bespoke evolutionary algorithm is specifically designed to minimise the number of simulations required, ensuring the tool is computationally practical. The effectiveness of the tool in this regard is validated via the derivation of exact convergence measures, for a tractable search space, and via comparisons with benchmark algorithms, for an intractable search space. Compared with traditional design practices, the proposed tool offers an automated approach capable of efficiently exploring a wide range of solutions, providing decision-makers with a set of optimal solutions from which they can make informed investment decisions. The presented methods provide a robust framework for optimising a variety of BGI features in complex urban environments.