Workflow decomposition algorithm for scheduling with quantum annealer-based hybrid solver

TL;DR

This paper presents the SPWD heuristic algorithm for decomposing large workflow scheduling problems, enabling efficient use of a quantum annealer-based hybrid solver for complex instances.

Contribution

The paper introduces the SPWD algorithm, a novel method for workflow decomposition that allows large problems to be solved with a quantum annealer-based hybrid solver.

Findings

SPWD enables scheduling of large workflows that exceed solver capacity.

The algorithm improves scalability of quantum annealer-based workflow scheduling.

Experimental results validate the effectiveness of the approach.

Abstract

We introduce the Series-Parallel Workflow Decomposition (SP\-WD) heuristic algorithm for the Workflow Scheduling Problem (WSP) decomposition. We demonstrate that the SPWD algorithm facilitates the scheduling of large WSP instances with the hybrid D-Wave Constrained Quadratic Model solver, enabling the scheduling of instances that would otherwise exceed its capacity limitations. We also describe the accompanying execution environment used to obtain the results of the experiments with real-life workflow instances available in the WfCommons standardization initiative repository.