Parallel genetic algorithms to solve dynamic task scheduling problems efficiently by taking into account the energy

Jia Luo 1
1 LAAS-CDA - Équipe Calcul Distribué et Asynchronisme
LAAS - Laboratoire d'analyse et d'architecture des systèmes
Abstract : Due to new government legislation, customers’ environmental concerns and continuously rising cost of energy, energy efficiency is becoming an essential parameter of industrial manufacturing processes in recent years. Most efforts considering energy issues in scheduling problems have focused on static scheduling. But in fact, scheduling problems are dynamic in the real world with uncertain new arrival jobs after the execution time. In this thesis, two energy efficient dynamic scheduling problems are studied. Model I analyzes the total tardiness and the makespan with a peak power limitation while considering the flexible flow shop with new arrival jobs. A periodic complete rescheduling approach is adopted to represent the optimization problem. Model II concerns an investigation into minimizing total tardiness and total energy consumption in the job shop with new urgent arrival jobs. An event driven schedule repair approach is utilized to deal with the u! pdated schedule. As an adequate renewed scheduling plan needs to be obtained in a short response time in dynamic environment, two parallel Genetic Algorithms (GAs) are proposed to solve these two models respectively. The parallel GA I is a CUDA-based hybrid model consisting of an island GA at the upper level and a fine-grained GA at the lower level. It combines metrics of two hierarchical layers and takes full advantage of CUDA’s compute capability. The parallel GA II is a dual heterogeneous design composed of a cellular GA and a pseudo GA. The islands with these two different structures increase the population diversity and can be well parallelized on GPUs simultaneously with multi-core CPU. Finally, numerical experiments are conducted and show that our approaches can not only solve the problems flexibly, but also gain competitive results and reduce time requirements.
Complete list of metadatas

Cited literature [60 references]  Display  Hide  Download

https://hal.laas.fr/tel-02009769
Contributor : Christine Fourcade <>
Submitted on : Wednesday, February 6, 2019 - 3:33:46 PM
Last modification on : Saturday, October 26, 2019 - 1:31:47 AM
Long-term archiving on : Tuesday, May 7, 2019 - 3:09:47 PM

File

 Restricted access
To satisfy the distribution rights of the publisher, the document is embargoed until : 2021-02-06

Please log in to resquest access to the document

Identifiers

  • HAL Id : tel-02009769, version 1

Citation

Jia Luo. Parallel genetic algorithms to solve dynamic task scheduling problems efficiently by taking into account the energy. Networking and Internet Architecture [cs.NI]. Université Toulouse 3 Paul Sabatier (UT3 Paul Sabatier), 2019. English. ⟨tel-02009769⟩

Share

Metrics

Record views

102