just for info: PyVRP v0.13.0+ does not use HGS anymore

[cyril23]

you wrote

Algorithms Implemented

1. Hybrid Genetic Search (HGS)
   Description: A state-of-the-art genetic algorithm that combines evolutionary operators with local search heuristics. HGS uses population-based search with crossover, mutation, and local improvement operators.

Implementation: Uses the pyVRP library, which provides an efficient HGS implementation.

and according to https://arxiv.org/pdf/2403.13795 PyVRP really did use HGS in the past - until not long ago.

Your current requirements.txt uses pyvrp==0.6.3 which still did use HGS - so your benchmarks really showed a performant HGS implementation.

Just be aware that because of PyVRP/PyVRP#533 starting with v0.13.0 PyVRP moved from HGS to ILS, see https://github.com/PyVRP/PyVRP/blob/v0.13.2/docs/source/setup/introduction_to_ils.rst

PyVRP provides a high-performance implementation of the iterated local search (ILS) algorithm for vehicle routing problems (VRPs). ILS is a single-trajectory algorithm that improves a solution by repeated applications of small perturbations and local improvement procedures. This approach effectively balances between exploration and exploitation of the search space.

Note

For a more thorough introduction to ILS for VRPs, we refer to the works of Lourenço et al. (2019) and Accorsi and Vigo (2021).

(hint: maybe you'd like to add a fifth state-of-the-art algorithm to your benchmarks, PyVRP 0.13.0+ and its ILS 👍 )

Anyhow your benchmarks looked interesting, so I wanted to let you know that.

[arnobt78]

Thank you so much for bringing this to my attention! 🙏 This is really valuable information, and I appreciate you taking the time to share it.

You're absolutely right - when I developed this project, I was using pyvrp==0.6.3, which indeed used HGS (Hybrid Genetic Search) at that time. The benchmarks and results in this repository reflect the HGS implementation from that version.

I wasn't aware that PyVRP moved from HGS to ILS (Iterated Local Search) starting with v0.13.0. Thank you for the heads up and for sharing the relevant links:

• PyVRP/PyVRP#533
• Introduction to ILS documentation

My Plan Going Forward

I'm planning to update this project to address this change:

1. Keep the current HGS implementation (pyvrp==0.6.3) as a fallback/reference implementation
2. Add the new ILS algorithm using PyVRP v0.13.0+ as a fifth algorithm in the comparison
3. Update the documentation to reflect both implementations and their differences
4. Expand the benchmarks to compare HGS vs ILS performance

Future Web Application

I'm also working on converting this project into a dynamic web application as a real-world demo project using:

• React for the frontend
• Python (Flask/FastAPI) for the backend
• Interactive UI for:
  • Dataset selection
  • Algorithm parameter tuning
  • Real-time visualization
  • Results comparison

In the web application, I plan to:

• Keep the old HGS implementation (pyvrp==0.6.3) as a fallback option
• Implement the new ILS algorithm (PyVRP v0.13.0+) as the primary/default option
• Allow users to choose between HGS and ILS implementations
• Run everything server-side with results displayed in the browser UI

This will make it easier for users to experiment with different algorithms and see the performance differences in real-time.

Next Steps

I'll create an issue to track this update and will:

• Update the codebase to support both HGS and ILS
• Add comprehensive documentation about the algorithm change
• Update benchmarks to include ILS results
• Ensure backward compatibility for existing users

Thanks again for the helpful information! This kind of community feedback is what makes open-source projects better. If you have any other suggestions or would like to contribute, feel free to reach out! 🚀

---

References:

• PyVRP Issue #533
• PyVRP ILS Introduction
• Lourenço et al. (2019) and Accorsi and Vigo (2021) - ILS for VRPs