An enhanced connected banking system optimizer with multiple strategies for numerical optimization problems.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Meta-heuristics

Background:

• Connected Banking System Optimizer (CBSO) is a parameter-free meta-heuristic for constrained optimization.
• CBSO exhibits limitations in inter-population communication and exploration-exploitation balance, leading to premature convergence.
• Inadequate search space coverage hinders CBSO's effectiveness on complex problems.

Purpose of the Study:

• To enhance the Connected Banking System Optimizer (CBSO) algorithm.
• To address CBSO's limitations in search space coverage and premature convergence.
• To develop a robust and efficient meta-heuristic for engineering constrained optimization problems.

Main Methods:

• Introduced an enhanced variant: ECBSO.
• Integrated feedback selection, regenerative population, and distribution estimation strategies.
• Conducted experiments on the CEC-2017 benchmark suite and real-world engineering problems.
• Employed statistical validation including Wilcoxon rank-sum, Friedman, and Nemenyi tests.

Main Results:

• ECBSO demonstrated superior optimization efficacy and robustness compared to existing algorithms.
• Achieved competitive average Friedman ranks across various dimensions (10D, 30D, 50D, 100D).
• Successfully solved ten real-world engineering constrained optimization problems with remarkable stability.

Conclusions:

• ECBSO effectively overcomes the limitations of CBSO, particularly premature convergence.
• The enhanced strategies significantly improve search space coverage and optimization performance.
• ECBSO is a robust and effective meta-heuristic variant suitable for complex engineering optimization tasks.