2025 | hadi e arbabi

an emergent optimal resource allocation for climate resilience of transport infrastructure networks

current infrastructure networks must be climate resilient to continue meeting service demand into the next decades with climate change rapidly pushing infrastructure assets towards or beyond their initial design envelope. at system level, this corresponds to the ability to deliver services when parts of the infrastructure become isolated following local asset failures. local shielding strategies are typically formulated using abstract network metrics or global optimization methods. the former are agnostic to the specificity of infrastructure systems, while the latter tend to be hardly scalable for large infrastructure networks. here, we develop an optimal limited resource allocation strategy to increase network resilience, combining the input sparsity of abstract network metrics with transparency of optimization methods. we focus on transport networks and maximizing the expected throughput of services. we consider upgrading costs as proportional to the desired increase in failure load from climate shocks. we benchmark our method by applying it to the uk freight railway considering shocks induced by an end of- century rcp8.5 climate change scenario. a closed form solution naturally emerges for the ranking of the network assets that allows for optimal distribution of limited asset reinforcement investments. we show that this attains better resilience improvements compared to existing heuristic global optimization methods.

references

papers

RSPA

An Emergent Optimal Resource Allocation for Climate Resilience of Transport Infrastructure Networks

Qianqian Li, Hadi Arbabi, and Giuliano Punzo

Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences, Jul 2025

Abs HTML PDF

Current infrastructure networks must be climate resilient to continue meeting service demand into the next decades with climate change rapidly pushing infrastructure assets towards or beyond their initial design envelope. At system level, this corresponds to the ability to deliver services when parts of the infrastructure become isolated following local asset failures. Local shielding strategies are typically formulated using abstract network metrics or global optimization methods. The former are agnostic to the specificity of infrastructure systems, while the latter tend to be hardly scalable for large infrastructure networks. Here, we develop an optimal limited-resource allocation strategy to increase network resilience, combining the input sparsity of abstract network metrics with transparency of optimization methods. We focus on transport networks and maximizing the expected throughput of services. We consider upgrading costs as proportional to the desired increase in failure load from climate shocks. We benchmark our method by applying it to the UK freight railway considering shocks induced by an end-of-century RCP8.5 climate change scenario. A closed-form solution naturally emerges for the ranking of the network assets that allows for optimal distribution of limited asset reinforcement investments. We show that this attains better resilience improvements compared to existing heuristic global optimization methods.