GPU-accelerated stochastic predictive control of drinking water networks

TL;DR

This paper introduces a GPU-accelerated stochastic model predictive control method for water networks, significantly reducing computational time by exploiting problem structure and parallelization, demonstrated on Barcelona's water system.

Contribution

It presents a novel GPU-based algorithm that efficiently solves stochastic predictive control problems by leveraging problem structure and parallel computing techniques.

Findings

Reduced computational time for stochastic control problems

Successful application to Barcelona's water network case study

Demonstrated scalability and efficiency of the GPU approach

Abstract

Despite the proven advantages of scenario-based stochastic model predictive control for the operational control of water networks, its applicability is limited by its considerable computational footprint. In this paper we fully exploit the structure of these problems and solve them using a proximal gradient algorithm parallelizing the involved operations. The proposed methodology is applied and validated on a case study: the water network of the city of Barcelona.