A Hybridization of Dragonfly Algorithm Optimization and Angle Modulation Mechanism for 0-1 Knapsack Problems.

Area of Science:

• Computational Intelligence
• Optimization Algorithms
• Metaheuristics

Background:

• The dragonfly algorithm (DA) is a nature-inspired metaheuristic effective for continuous optimization.
• Adapting DA for binary optimization problems requires novel mechanisms.
• Existing binary adaptations may lack sufficient stability or convergence speed.

Purpose of the Study:

• To adapt the dragonfly algorithm (DA) for binary optimization problems.
• To enhance the stability and convergence speed of the binary DA.
• To evaluate the performance of the proposed algorithms on benchmark and engineering problems.

Main Methods:

• Introduction of an angle modulation mechanism (AMDA) to generate binary solutions from DA.
• Development of an improved AMDA (IAMDA) by adding a coefficient for vertical displacement adjustment.
• Testing AMDA and IAMDA on 12 zero-one knapsack problems and 13 classic benchmark functions.

Main Results:

• IAMDA shows a superior convergence speed compared to AMDA and other tested algorithms.
• IAMDA achieves higher solution quality in solving zero-one knapsack and benchmark problems.
• The angle modulation mechanism effectively translates DA's continuous optimization capabilities to the binary domain.

Conclusions:

• The proposed IAMDA algorithm offers an effective and efficient approach for binary optimization.
• IAMDA provides a significant improvement over the original AMDA in terms of speed and solution quality.
• This work extends the applicability of the dragonfly algorithm to a wider range of discrete optimization tasks.