Hybrid leader based optimization: a new stochastic optimization algorithm for solving optimization applications.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Metaheuristics

Background:

• Optimization problems are prevalent across scientific and engineering disciplines.
• Existing algorithms often struggle with balancing exploration and exploitation.
• Developing efficient and robust optimization techniques is crucial.

Purpose of the Study:

• Introduce a novel optimization algorithm, Hybrid Leader-based Optimization (HLBO).
• Mathematically model HLBO's exploration and exploitation phases.
• Evaluate HLBO's effectiveness on diverse benchmark functions and engineering problems.

Main Methods:

• HLBO guides algorithm population using a hybrid leader.
• Algorithm stages are modeled mathematically in two phases: exploration and exploitation.
• Tested on 23 standard unimodal/multimodal benchmark functions and IEEE CEC 2017 functions.

Main Results:

• HLBO shows high exploitation ability for unimodal functions, converging to global optima.
• HLBO demonstrates high exploration ability for multimodal functions, scanning search spaces effectively.
• HLBO outperforms ten well-known algorithms in convergence and global solution finding.

Conclusions:

• HLBO is efficient in handling complex objective functions.
• HLBO proves applicable to real-world engineering design problems.
• HLBO offers a superior approach to optimization compared to existing methods.