Scheduling under Precedence, Communication, and Energy Constraints

TL;DR

This paper develops scheduling algorithms for multiple processors that optimize task completion time while respecting energy limits and task dependencies, extending traditional models to include energy constraints.

Contribution

It introduces convex programming algorithms that incorporate energy constraints into scheduling, matching the performance of existing methods without energy considerations.

Findings

Algorithms effectively handle energy bounds in scheduling.

Performance comparable to traditional scheduling algorithms.

Provides a framework for energy-aware scheduling in multi-processor systems.

Abstract

We consider the problem of scheduling a set of $n$ tasks on $m$ processors under precedence, communication, and global system energy constraints to minimize makespan. We extend existing scheduling models to account for energy usage and give convex programming algorithms that yield essentially the same results as existing algorithms that do not consider energy, while adhering to a strict energy bound.