This paper discusses the issue of niche formation in genetic algorithms (GAs) for multimodal optimization. Niche formation allows evolutionary algorithms to maintain multiple solutions in different regions of the phenotypic space, which is essential for multiobjective optimization and complex system simulation. However, conventional GAs tend to group the final population around the fittest individual, leading to loss of diversity. Niching methods help maintain solutions in different regions of interest. The paper presents a crowding method for niche formation and analyzes its performance on two multimodal functions. Three distance criteria are used: phenotypic, genotypic, and fitness. Phenotypic distance measures the difference in the actual values of the solutions, genotypic distance compares the binary strings, and fitness distance measures the difference in fitness values. The study shows that the phenotypic criterion performs best in maintaining diversity and finding optimal solutions. For function f1 with equal height peaks, the phenotypic criterion provided the best results in terms of optimal hits and minimum Ebest. For function f2 with decreasing height peaks, the phenotypic criterion also performed best, with niche counts nearly proportional to peak heights. The results indicate that the phenotypic criterion is the most convenient approach for niche formation in multimodal optimization. The paper concludes that the phenotypic criterion is more effective in maintaining diversity and finding optimal solutions compared to the other two criteria. Further research is needed to explore variants of crowding that can regain diversity when it is lost.This paper discusses the issue of niche formation in genetic algorithms (GAs) for multimodal optimization. Niche formation allows evolutionary algorithms to maintain multiple solutions in different regions of the phenotypic space, which is essential for multiobjective optimization and complex system simulation. However, conventional GAs tend to group the final population around the fittest individual, leading to loss of diversity. Niching methods help maintain solutions in different regions of interest. The paper presents a crowding method for niche formation and analyzes its performance on two multimodal functions. Three distance criteria are used: phenotypic, genotypic, and fitness. Phenotypic distance measures the difference in the actual values of the solutions, genotypic distance compares the binary strings, and fitness distance measures the difference in fitness values. The study shows that the phenotypic criterion performs best in maintaining diversity and finding optimal solutions. For function f1 with equal height peaks, the phenotypic criterion provided the best results in terms of optimal hits and minimum Ebest. For function f2 with decreasing height peaks, the phenotypic criterion also performed best, with niche counts nearly proportional to peak heights. The results indicate that the phenotypic criterion is the most convenient approach for niche formation in multimodal optimization. The paper concludes that the phenotypic criterion is more effective in maintaining diversity and finding optimal solutions compared to the other two criteria. Further research is needed to explore variants of crowding that can regain diversity when it is lost.