Gaussian Quantum Bat Algorithm with Direction of Mean Best Position for Numerical Function Optimization.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Metaheuristics

Background:

• The Quantum-behaved Bat Algorithm (QMBA) is effective but can suffer from premature convergence due to uniform probability distributions.
• A limited search range in QMBA hinders its ability to escape local optima.

Purpose of the Study:

• To propose a novel Gaussian quantum bat algorithm with mean best position directed (GQMBA).
• To enhance the QMBA's ability to avoid premature convergence and improve search performance.
• To apply GQMBA to solve numerical function optimization problems.

Main Methods:

• The Gaussian quantum bat algorithm (GQMBA) was developed by incorporating Gaussian probability distribution into the QMBA framework.
• Random number sequences for stochastic coefficients were generated using Gaussian distribution instead of uniform distribution.
• GQMBA's performance was evaluated on nineteen benchmark functions and compared against other optimization algorithms.

Main Results:

• The proposed GQMBA algorithm demonstrated superior search performance compared to other algorithms across most benchmark functions.
• The application of Gaussian distribution effectively mitigated premature convergence issues observed in the original QMBA.
• GQMBA showed improved accuracy and robustness in solving numerical function optimization tasks.

Conclusions:

• The Gaussian quantum bat algorithm (GQMBA) is a promising advancement over existing QMBA methods.
• Gaussian probability distribution is an effective technique for enhancing bat algorithm performance and avoiding local optima.
• GQMBA offers a more robust and accurate approach for numerical function optimization.