optimization-research-workflow

An optimization research workflow connects theory framing, Julia implementation, parameter tuning, benchmarking, and publication. Using a template-driven approach with customizable context files ensures reproducible experiments across multiple problems and configurations.

To start a new optimization study, copy the workflow template to your project directory, then tailor the CLAUDE.md and jcode/CLAUDE.md files to match your specific algorithm requirements and configuration parameters.

DataFrames and CSV are required dependencies for handling data within the optimization benchmarking workflow. ProgressMeter is also included to track the execution status of parameter tuning and experiment phases.

Yes, the workflow supports parameter tuning and phase transitions across multiple problems and configurations. Reusable templates and standard workflow references accelerate project startup while maintaining experiment reproducibility.

Template-driven benchmarking improves reproducibility by standardizing the connection between theory, implementation, and publication steps. Reusable references ensure consistent experiment configurations and structured outputs across different optimization studies.

Standard workflow templates require explicit context configuration via CLAUDE.md and jcode/CLAUDE.md files. Users must customize these implementation guides to match their specific algorithm, as the template alone does not automatically generate optimization logic.