Experimental Data Sets for the study "Benchmarking a $(\mu+\lambda)$ Genetic Algorithm wit

Leiden University researchers Furong Ye and Carola Doerr produced experimental results for benchmarking a novel (μ+λ) genetic algorithm. The dataset includes results from 25 pseudo-Boolean problems from IOHprofiler and detailed runs on OneMax and LeadingOnes problems with varying parameters. It contains raw running times and summary statistics from 100 independent runs per configuration.

Use Cases

• Benchmarking new genetic algorithm variants based on performance data across 25 pseudo-Boolean problems.
• Analyzing the impact of crossover probability (p_c) on algorithm performance based on the described experimental setup.
• Comparing the effectiveness of standard bit mutation versus fast mutation operators in a (μ+λ) GA.
• Studying the fixed-target running time and success rates for OneMax and LeadingOnes problems across different problem sizes (n) and population parameters.

Strengths

• Results from 25 distinct pseudo-Boolean benchmark problems sourced from IOHprofiler.
• Includes data from 100 independent runs for OneMax and LeadingOnes problems, providing statistical robustness.
• Covers a wide parameter space, including multiple values for μ, λ, and p_c as described.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

Leiden University

Collection Method

Experimental results from running a proposed (μ+λ) genetic algorithm on benchmark problems.