An Improved Whale Migration Optimization Algorithm for Cooperative UAV 3D Path Planning.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Artificial Intelligence

Background:

• The original Whale Migration Algorithm (WMA) faces challenges with premature convergence and insufficient local exploitation in high-dimensional multimodal optimization problems.
• Addressing these limitations is crucial for developing more robust and efficient optimization techniques.

Purpose of the Study:

• To introduce an Improved Whale Migration Algorithm (IWMA) that overcomes the shortcomings of the WMA.
• To enhance the algorithm's performance in terms of population diversity, exploration-exploitation balance, global escaping, and local refinement.

Main Methods:

• IWMA incorporates circle chaotic initialization for improved population diversity.
• A three-layer cooperative search framework is implemented to balance exploration and exploitation.
• A dynamic adaptive mechanism with t-distribution re-exploration is utilized for global escaping and local refinement.

Main Results:

• IWMA demonstrated superior performance on the CEC2017 benchmark suite, outperforming seven other algorithms on most functions across 30 and 50 dimensions.
• In multi-UAV path planning, IWMA reduced average cost by 14.5% compared to WMA and achieved significant reductions against seven competitors.
• Statistical tests and visualizations confirmed the effectiveness of the proposed IWMA enhancements.

Conclusions:

• The proposed enhancements to the Whale Migration Algorithm are effective, leading to improved performance in both benchmark optimization and practical multi-UAV path planning.
• IWMA consistently produces safer and smoother trajectories in complex multi-UAV cooperative mission planning scenarios.
• The study validates the successful transfer of performance gains from theoretical benchmarks to real-world applications.