A Modified Reptile Search Algorithm for Numerical Optimization Problems.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Swarm Intelligence

Background:

• The Reptile Search Algorithm (RSA) is a swarm-based metaheuristic inspired by crocodile hunting behaviors.
• Existing RSA variants face challenges including low population diversity, suboptimal exploration-exploitation balance, and premature convergence to local optima.

Purpose of the Study:

• To introduce a Modified Reptile Search Algorithm (MRSA) that addresses the limitations of the standard RSA.
• To enhance the performance of metaheuristic optimization through improved diversity and exploration-exploitation dynamics.

Main Methods:

• An adaptive chaotic reverse learning strategy was implemented to increase population diversity.
• An elite alternative pooling strategy was developed to achieve a better balance between exploitation and exploration.
• A shifted distribution estimation strategy was incorporated to refine the evolutionary direction and boost overall performance.

Main Results:

• MRSA demonstrated superior performance across 23 benchmark functions and IEEE CEC2017 benchmark functions.
• The algorithm showed significant improvements in convergence accuracy, speed, and stability in robot path planning problems.
• Statistical tests (Friedman, Wilcoxon signed-rank) confirmed MRSA's outperformance compared to other algorithms.

Conclusions:

• The proposed MRSA effectively overcomes the limitations of the standard RSA, particularly in population diversity and exploration-exploitation balance.
• MRSA offers a robust and efficient alternative for complex optimization tasks, including path planning.
• The enhancements contribute to more accurate, faster, and stable optimization outcomes.