An enhanced escape algorithm with comprehensive learning and Cauchy-Gaussian mutation for reservoir optimization

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Engineering Applications

Background:

• Global optimization of high-dimensional problems is challenging due to premature convergence and diversity loss.
• Existing metaheuristics often struggle to balance exploration and exploitation effectively.

Purpose of the Study:

• To introduce CLGMESC, an enhanced Escape Algorithm (ESC) designed to address limitations in global optimization.
• To improve population diversity and facilitate escape from local optima in complex landscapes.

Main Methods:

• CLGMESC integrates a dimension-wise comprehensive learning (CL) strategy for stagnant individuals.
• A hybrid Cauchy-Gaussian mutation (HCGM) operator with adaptive weighting balances exploration and exploitation.
• Evaluations were conducted on the CEC2017 benchmark suite and a reservoir production optimization problem.

Main Results:

• CLGMESC ranked first among ten advanced metaheuristics on the CEC2017 benchmark suite.
• Statistical tests confirmed CLGMESC's superiority across most test functions.
• In reservoir optimization, CLGMESC achieved the highest Net Present Value with the lowest standard deviation.

Conclusions:

• CLGMESC demonstrates superior performance in global optimization tasks compared to existing methods.
• The algorithm effectively maintains exploration-exploitation balance and escapes local optima.
• CLGMESC is a reliable and robust solution for computationally intensive real-world engineering problems.