Improved Zebra Optimization Algorithm with Multi Strategy Fusion and Its Application in Robot Path Planning.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Metaheuristics

Background:

• The baseline Zebra Optimization Algorithm (ZOA) suffers from premature convergence and local optima trapping.
• Existing metaheuristics often struggle with balancing exploration and exploitation effectively.
• Need for robust algorithms to handle complex, real-world optimization problems.

Purpose of the Study:

• To develop an improved optimization algorithm, the Multi-Strategy Enhanced Zebra Optimization Algorithm (MZOA).
• To address the limitations of the standard ZOA, enhancing its convergence and diversity.
• To validate the effectiveness of MZOA against contemporary metaheuristics on benchmark and application-specific problems.

Main Methods:

• Integration of triangular walk operators for balanced exploration and exploitation.
• Incorporation of Levy flight mechanisms for enhanced solution space traversal.
• Implementation of lens imaging inversion learning to boost population diversity and prevent stagnation.

Main Results:

• MZOA demonstrated a 15.8% performance improvement over the basic ZOA on CEC2005 and CEC2017 benchmark suites.
• Significant reduction in robot path planning by 8.7% compared to the basic ZOA.
• Consistent achievement of optimal solutions across diverse and complex optimization scenarios.

Conclusions:

• The MZOA effectively overcomes the limitations of the ZOA, offering enhanced accuracy and reliability.
• MZOA proves to be a robust computational tool for complex optimization challenges.
• The algorithm shows practical effectiveness and operational reliability in both synthetic and real-world applications.