Quad-Criteria Task Scheduling in Multicores Based on NSGAIII

Wang，Jiaxuan1,2; Tan，Zheng1,3; Yuan，Bo1

10.1007/978-3-031-36622-2_19

2023

0302-9743

1611-3349

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

13968 LNCS

229-241

It is very challenging to schedule multiple tasks in multicore systems, as there are multiple antagonistic optimization objectives, such as makespan, energy consumption, reliability, and peak temperature. To produce Pareto front in this four-dimensional space, the previous ϵ- constraint method transforms the original multi-objective optimization problem into multiple constrained single objective optimization problems, which relies heavily on the thresholds chosen for the constraints. As population-based global optimization methods, Multi-Objective Evolutionary Algorithms (MOEA) have been successfully applied in solving real-world multi-objective scheduling problems. In this work, we propose a problem-specific encoding strategy for the quad-criteria task scheduling problem and use the Non-dominated Sorting Genetic Algorithm-III (NSGAIII) as the search engine. Compared with the ϵ- constraint method, experimental results over a large set of benchmarks demonstrate that the proposed NSGAIII based method can provide a better solution set within approximately the same running time, especially when the number of cores is large.

Evolutionary algorithm Multi-objective optimization Multicore systems Task scheduling

First ; Corresponding

English

2-s2.0-85169025508

Scopus

1.Department of Computer Science and Engineering,Southern University of Science and Technology,Shenzhen,China
2.Academy for Advanced Interdisciplinary Studies,Southern University of Science and Technology,Shenzhen,China
3.School of Computer Science,University of Birmingham,Birmingham,United Kingdom

Wang，Jiaxuan,Tan，Zheng,Yuan，Bo. Quad-Criteria Task Scheduling in Multicores Based on NSGAIII[C],2023:229-241.