A Dimensionality-Reduction Strategy to Compute Shortest Paths in Urban Water Networks
The efficient computation of shortest paths in complex networks is essential to face new challenges related to critical infrastructures such as a near real-time monitoring and control and the management of big size systems. In particular, using information on the minimum paths in water distribution networks (WDNs) allows to track the diffusion of contaminants and to quantify the resilience and criticality of the system. This is, ultimately, approached by considering dynamically changing path-weights that depend on the flow or on other information available at run-time. These analyses tipically include all the WDN assets but reducing the high degree of physical details with a minimum lost of key information for their performance assessment. This paper proposes a strategy to compute minimum paths that is based on a dimensionality-reduction process. Specifically, the network is partitioned into communities and suitably modified to obtain a reduced complexity representation (e.g., in terms of number of nodes and links). The paper shows how this novel, reduced representation is equivalent to the traditional network on computing the shortest paths. The proposed approach is validated considering two utility networks as case studies. The results show that the proposed method provides the exact solution for the shortest path with a computational-time reduction consistently over 50\% and up to 90\% for some cases. Furthermore, the application of the proposal on WDNs partitioning shows both hydraulic and economic advantages thanks to their monitoring and controlling at near real-time.