CORDIS · FullScaleOPT

Fully Scalable Network Optimization

Code

FullScaleOPT

Programme

HORIZON

Call

ERC-2025-ADG

Status

SIGNED

Period

2026-08-01 – 2031-07-31

Official data · CORDIS · synced 2026-10-18 03:29 · official record

Participants

Description

Networks are ubiquitous and integral throughout society, sciences, and everyday life. Examples include utility, transportation, and communication networks as well as abstract networks used across all sciences and engineering to model a wide variety of relations, dependencies, and dynamics. Scalable parallel network optimization algorithms are crucial for optimizing, understanding, operating, and analyzing these networks efficiently. The exploding scale and exponential growth of modern networks and optimization tasks put increasingly strict and urgent requirements on algorithms to scale both up and down, i.e., (i) the total work and computational resources required must scale up at most linearly with the size of the network task, and (ii) algorithms must parallelize, that is, running times must scale down linearly with the number of processors or machines put to work.Despite this growing significance, long-standing fundamental barriers in parallel computing place full scalability outside the current state of the art and seemingly out of reach for all existing algorithm design techniques. This applies even to the most basic tasks, including breadth-first search and directed connectivity, making it one of the most pressing and consequential open problems in the foundations of parallel computing.The goal of this project is to attack this fundamental challenge head-on and build the mathematical and algorithmic tools for achieving fully scalable network optimization algorithms for many fundamental tasks, particularly network flows. The proposal gives a detailed research program and strong evidence that the proposed innovative ideas and methods can overcome the barriers in the area and lead to the strongest possible form of scalability for many practically important and pervasive network optimization tasks.

Official data · CORDIS · synced 2026-10-18 03:29 · official record · Licence: © European Union, CC BY 4.0 (Commission reuse policy).