Version 0.7.0: Multi-Objective Optimization, New Solvers, and Major Usability Upgrades

Version 0.7.0: Multi-Objective Optimization, New Solvers, and Major Usability Upgrades

This is a major feature release that transforms EPTHenOpt from a set of scripts into a professional, installable Python package with significantly expanded optimization capabilities. The focus of this release is on providing more powerful tools for real-world engineering analysis, including multi-objective optimization and a wider selection of algorithms.

✨ New Features

1. Multi-Objective Optimization with NSGA-II

The most significant new feature is the introduction of multi-objective optimization to find the optimal trade-off between economic cost and environmental impact.

• Dual Objectives: The package can now simultaneously optimize for Total Annualized Cost (TAC) and CO₂ Emissions.
• New Solver: This is powered by the industry-standard NSGA-II (Non-dominated Sorting Genetic Algorithm II).
• Pareto Front Export: When running in multi-objective mode, the entire Pareto front of non-dominated solutions is exported to pareto_front_results.csv, perfect for plotting and analysis in Excel.
• Usage: Activate this mode with the --objective multi command-line flag.

2. Expanded Suite of Optimization Algorithms

The package is no longer limited to GA and TLBO. Four powerful metaheuristics have been added to the library:

• Particle Swarm Optimization (PSO): A swarm intelligence algorithm inspired by the flocking behavior of birds.
• Simulated Annealing (SA): A probabilistic algorithm excellent for escaping local optima.
• Ant Colony Optimization (ACO): A constructive metaheuristic that models how ants find the shortest path using pheromone trails.
• NSGA-II: The new default for multi-objective runs.

3. JSON-Based Configuration

• You can now manage all run parameters from a central config.json file. This is ideal for managing complex setups and ensuring reproducibility.
• Command-line arguments will always override settings from the JSON file, giving you the best of both worlds.

4. Structured Results Export

• A new --output_dir argument allows you to save all results from a run into a specified directory.
• The output includes:
  • summary.json: A complete cost breakdown.
  • network_structure.csv: A detailed list of all exchangers, their duties, areas, and temperatures.
  • stream_results.csv: A summary of the final state of all process streams and whether they met their targets.

🚀 Improvements & Changes

• Full Python Package: EPTHenOpt is now a proper, installable Python package with a setup.py. You can install it locally with pip install -e ..
• New Command-Line Interface: The package now provides a clean command-line entry point: run_hen_problem.
• New Programmatic API: You can now easily import and use the package in your own Python scripts via the EPTHenOpt.run() function.
• Detailed Costing Model: The config.json and internal models now support separate default cost parameters for process exchangers, heaters, and coolers, allowing for more realistic problem definitions.
• Enhanced Documentation:
  • A new LITERATURE.md file explains the theoretical background of the Stage-Wise Superstructure model.
  • A new CONTRIBUTING.md file provides guidelines for developers.
  • The project is now set up for automated API documentation generation using Sphinx.

How to Install/Upgrade

To install this new version, clone the repository and run the pip installation command from the project's root directory. Using editable mode is recommended for development.

pip install -e '.[dev]'