Multi‑strategy Equilibrium Optimizer: An improved meta-heuristic tested on numerical optimization and engineering...

Area of Science:

• Computational Intelligence
• Optimization Algorithms

Background:

• The Equilibrium Optimizer (EO) is a novel algorithm for global optimization.
• EO can get stuck in local optima for complex numerical and engineering problems, reducing accuracy.

Purpose of the Study:

• To propose an improved Equilibrium Optimizer (IEO) addressing the limitations of the original EO.
• To enhance the accuracy, robustness, and efficiency of the optimization process.

Main Methods:

• Introduced Tent mapping for diverse initial population distribution.
• Implemented a nonlinear time parameter to balance exploration and exploitation.
• Incorporated Lens Opposition-based Learning (LOBL) to improve population diversity and avoid local optima.

Main Results:

• IEO demonstrated superior performance on 23 classical, IEEE CEC2017, and IEEE CEC2019 benchmark problems.
• Statistical analysis (Friedman test, box plots) confirmed the algorithm's stability and robustness.
• IEO achieved high optimization efficiency and cost minimization on six engineering design problems.

Conclusions:

• The proposed IEO effectively overcomes local optima and improves solution accuracy.
• IEO exhibits excellent robustness and efficiency for complex optimization tasks.
• The multi-strategy approach significantly enhances the capabilities of the Equilibrium Optimizer.