A novel link-based Multi-objective Grey Wolf Optimizer for Appliances Energy Scheduling Problem

Document Type

Article

Publication Title

Cluster Computing

Abstract

In this paper, a modified version of the Multi-objective Grey Wolf Optimizer (MGWO), known as linked-based GWO (LMGWO), is proposed for the Appliances Energy Scheduling Problem (AESP). The proposed LMGWO is utilized by combining the MGWO searching mechanism with a novel strategy, called neighbourhood selection strategy, to improve local exploitation capabilities. AESP is a problem that can be tackled by searching for the best appliances schedule according to a set of constraints and a dynamic pricing scheme(s) utilized for optimizing energy consumed at a particular period. Three objectives are considered to handle AESP: improving user comfort while reducing electricity bills and maintaining power systems’ performance. Therefore, AESP is modelled as a multi-objective optimization problem to handle all objectives simultaneously. In the evaluation results, the LMGWO is tested using a new dataset containing 30 power consumption scenarios with up to 36 appliances. For comparative purposes, the same linked-based neighbourhood selection strategy is utilized with other three optimization algorithms, including particle swarm optimization, salp swarm optimization, and wind-driven algorithm. The performance of the modified versions is compared with each other and that of the original versions to show their improvements. Furthermore, the proposed LMGWO is compared with eight state-of-the-art methods using their recommended datasets to show the viability of the proposed LMGWO. Interestingly, the proposed LMGWO is able to outperform the compared methods in almost all produced results.

First Page

4355

Last Page

4382

DOI

10.1007/s10586-022-03675-3

Publication Date

7-22-2022

Keywords

Appliances Energy Scheduling Problem, Linked Based Grey Wolf Optimizer, Multi-objective Grey Wolf Optimizer, Neighbourhood Selection Strategy

Comments

IR Deposit conditions:

OA version (pathway a) Accepted version

12 month embargo

Published source must be acknowledged

Must link to publisher version with DOI

Post-prints are subject to Springer Nature re-use terms

Share

COinS