Foundations for Workflow Application Scheduling on D-Wave System

Foundations for Workflow Application Scheduling on D-Wave System

Dawid Tomasiewicz, Maciej Pawlik, Maciej Malawski, Katarzyna Rycerz

In Valeria V. Krzhizhanovskaya, Gábor Závodszky, Michael H. Lees, Jack J. Dongarra, Peter M. A. Sloot, Sérgio Brissos & João Teixeira: Computational Science – ICCS 2020. 1 516--530. Cham: Springer International Publishing https://doi.org/10.1007/978-3-030-50433-5_40. ISBN: 978-3-030-50433-5

Many scientific processes and applications can be represented in the standardized form of workflows. One of the key challenges related to managing and executing workflows is scheduling. As an NP-hard problem with exponential complexity it imposes limitations on the size of practically solvable problems. In this paper, we present a solution to the challenge of scheduling workflow applications with the help of the D-Wave quantum annealer. To the best of our knowledge, there is no other work directly addressing workflow scheduling using quantum computing. Our solution includes transformation into a Quadratic Unconstrained Binary Optimization (QUBO) problem and discussion of experimental results, as well as possible applications of the solution. For our experiments we choose four problem instances small enough to fit into the annealer’s architecture. For two of our instances the quantum annealer finds the global optimum for scheduling. We thus show that it is possible to solve such problems with the help of the D-Wave machine and discuss the limitations of this approach.