An algorithmic framework for multiobjective optimization.

Area of Science:

• Computational Science
• Optimization Theory
• Algorithm Design

Background:

• Multiobjective (MO) optimization is crucial across many global fields.
• Existing metaheuristic techniques (DE, GA, GSA, PSO) and scalarization methods (weighted sum, NBI) face challenges with multiple objectives and computational overhead.
• Hybrid algorithms often introduce significant computational burdens.

Purpose of the Study:

• To propose an alternative framework for generating efficient and effective multiobjective optimization algorithms.
• To address the limitations of current methods, particularly for problems with more than two objectives.
• To minimize computational overhead in high-performance algorithm development.

Main Methods:

• Developing a novel framework leveraging algorithmic concepts tied to problem structure.
• Focusing on generating algorithms that are both efficient and effective.
• Prioritizing minimal computational overhead in the algorithm design process.

Main Results:

• A new framework for constructing multiobjective optimization algorithms is presented.
• The framework aims to enhance performance while reducing computational demands.
• It offers a structured approach to algorithm generation based on problem characteristics.

Conclusions:

• The proposed framework provides a promising alternative for tackling complex multiobjective optimization problems.
• It offers a pathway to developing high-performance algorithms with reduced computational costs.
• This approach is particularly relevant for scenarios involving numerous objectives and efficiency constraints.