A multi-strategy improved lemur optimiser for global optimisation and engineering applications.

Area of Science:

• Computational intelligence
• Optimization algorithms
• Swarm intelligence

Background:

• Existing Lemur Optimiser (LO) suffers from slow convergence and imbalanced exploration/exploitation.
• Need for advanced algorithms to address complex multi-objective and real-world optimization challenges.

Purpose of the Study:

• Propose a Multi-strategy Improved Lemur Optimiser (MILO) algorithm.
• Address the limitations of the original LO algorithm.
• Enhance performance in terms of convergence speed, accuracy, and stability.

Main Methods:

• Integration of Chebyshev chaotic mapping for initial population diversity.
• Implementation of a multielite-guided differential population evolution strategy to avoid local optima.
• Application of longitudinal and transversal crossover strategies for comprehensive solution space exploration.

Main Results:

• MILO significantly outperforms eleven established swarm intelligence optimisers on CEC2005, CEC2017, and CEC2022 benchmark functions.
• Demonstrated improvements in solution accuracy by 19.56%, 25.79%, and 45.73% for typical engineering design problems.
• MILO shows enhanced stability and accuracy in later evolutionary stages.

Conclusions:

• MILO effectively tackles flaws of the LO algorithm, offering improved convergence and exploration-exploitation balance.
• The proposed algorithm shows strong potential for solving complex multi-objective and real-world optimization problems.
• MILO represents a significant advancement in population evolution-based optimization techniques.