QL-ADIFA: Hybrid optimization using Q-learning and an adaptive logarithmic spiral-levy firefly algorithm.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Swarm Intelligence

Background:

• Optimization problems are prevalent across engineering and science.
• Meta-heuristics, like the firefly algorithm (FA), are effective for solving complex optimization tasks.
• Existing FA enhancements still have limitations in performance.

Purpose of the Study:

• To introduce an improved firefly algorithm, Q-learning based adaptive logarithmic spiral-Levy flight firefly algorithm (QL-ADIFA).
• To address the deficiencies of existing firefly algorithms.
• To enhance the exploration and exploitation capabilities of meta-heuristic optimization.

Main Methods:

• Integration of Q-learning with an enhanced firefly algorithm incorporating adaptive logarithmic spiral and Levy flight.
• Leveraging Q-learning for environmental awareness and memory in the firefly algorithm.
• Extensive numerical experiments on benchmark and engineering optimization problems.

Main Results:

• QL-ADIFA significantly outperforms existing meta-heuristic methods.
• Demonstrated effectiveness on 15 benchmark optimization functions.
• Successful application to 12 engineering problems, including cantilever arm, pressure vessel, three-bar truss, and CEC2020 constrained problems.

Conclusions:

• QL-ADIFA represents a substantial advancement in meta-heuristic optimization.
• The integration of Q-learning enhances the adaptive capabilities and memory of the firefly algorithm.
• The proposed method offers a robust and efficient solution for diverse engineering and scientific optimization challenges.