Comparative assessment of differently randomized accelerated particle swarm optimization and squirrel search...

Area of Science:

• Computational Intelligence
• Electrical Engineering
• Optimization Algorithms

Background:

• Nature-inspired metaheuristic algorithms are valuable when initial guesses are unknown.
• Random initialization is crucial for the effective deployment of these algorithms.
• Selective Harmonics Elimination (SHE) requires robust optimization techniques.

Purpose of the Study:

• To analyze the impact of different randomization types on metaheuristic algorithm performance.
• To evaluate how randomization affects solutions in Selective Harmonics Elimination (SHE).
• To compare the efficacy of various randomizations within Accelerated Particle Swarm Optimization (APSO) and Squirrel Search Algorithm (SSA).

Main Methods:

• Applied five distinct randomization types (exponential, normal, Rayleigh, uniform, Weibull) to APSO and SSA.
• Utilized these algorithms for solving the Selective Harmonics Elimination (SHE) problem.
• Conducted statistical analysis to assess the influence of randomization.

Main Results:

• The type of randomization applied demonstrably impacts algorithm operation.
• Different randomizations lead to variations in the fittest objective function values.
• The choice of randomization influences the quality of solutions obtained for SHE.

Conclusions:

• Randomization strategy is a critical factor in the success of metaheuristic optimization for SHE.
• Algorithm performance and solution quality are sensitive to the specific random distribution used.
• Further research into optimal randomization for specific problems is warranted.