Crossover versus mutation: a comparative analysis of the evolutionary strategy of genetic algorithms applied to...

Area of Science:

• Computer Science
• Operations Research
• Artificial Intelligence

Background:

• Genetic Algorithms (GAs) are widely used for combinatorial optimization.
• Research on GAs is extensive, but the impact of blind crossover operators is understudied.
• Combinatorial optimization problems are prevalent in various scientific and industrial domains.

Purpose of the Study:

• To objectively analyze the influence of blind crossover operators in GAs for combinatorial optimization.
• To compare the performance of GAs with different configurations against mutation-focused evolutionary algorithms.
• To provide a comprehensive understanding of operator selection in evolutionary computation.

Main Methods:

• Comparison of nine techniques: six Genetic Algorithm (GA) configurations and three mutation-focused evolutionary algorithms.
• Application of these techniques to four well-known combinatorial optimization problems.
• Statistical analysis of results using the normal distribution z-test for reliable comparison.

Main Results:

• Certain GA configurations demonstrated superior performance compared to mutation-only evolutionary algorithms.
• The study identified specific blind crossover operator strategies that are more effective for certain problem types.
• Statistical analysis confirmed significant differences in the performance of the evaluated techniques.

Conclusions:

• The choice of crossover operator significantly impacts GA performance in combinatorial optimization.
• Blind crossover operators can be effective, but their efficacy is problem-dependent.
• This research provides valuable data for selecting appropriate GA components for specific optimization tasks.