A fast and efficient teaching-learning-based optimization algorithm for large-scale multi-objective optimization problems

ABSTRACT

Multi-objective optimization problems with large-scale decision variables, known as LSMOPs, are characterized by their large-scale search space and multiple conflicting objectives to be optimized. They have been involved in many emergent real-world applications, such as feature and instance selection, data clustering, adversarial attack, vehicle routing problem, and others. In this work, we propose a new teaching-learning-based optimization algorithm to effectively tackle large-scale multi-objective optimization problems. The proposed approach ensures both the diversity of solutions, requested for high dimensionality of LSMOPs, and the balance between the exploitation and exploration during the optimization process. The experimental studies conducted on 54 instances of a large-scale multi-objective optimization problems test suite and the comparisons against well-known and state-of-the-art algorithms have shown the superiority of the proposed algorithm in terms of Pareto front approximation and computation time by using limited evaluation budget.

KEYWORDS:

• Teaching-learning-based optimization
• multi-objective optimization
• large-scale optimization

AMS 2010 MATHEMATICS SUBJECT CLASSIFICATION CODES::

• 90C06
• 90C29

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding and competing interests

The authors have no competing interests to declare that are relevant to the content of this article.

Data and code availability

The code and the datasets generated during and/or analyzed during the current study are available from the corresponding author on request.