Skip to content

Releases: Zhengzhong-You/RouteOpt

RouteOpt 2.0.0

11 Mar 06:03
@Zhengzhong-You Zhengzhong-You
v2.0.0
0d74406
Compare
Choose a tag to compare
RouteOpt 2.0.0 Latest
Latest

RouteOpt 2.0

We are excited to announce the release of RouteOpt 2.0, a major milestone that further enhances the capabilities of this comprehensive framework for solving Vehicle Routing Problems (VRPs) exactly. Built upon leading commercial solvers such as Gurobi, CPLEX, and MindOpt, RouteOpt 2.0 offers a refined modular structure, improved performance, and more streamlined workflows.

What is RouteOpt?

RouteOpt is a versatile toolkit designed for exact solutions to complex routing and optimization problems, particularly VRPs. It includes:

  • Rank-1 Cuts: A robust set of tools for cut generation, separation, reduced-cost updates in labeling, coefficient calculations, and more.
  • Rounded Capacity Cuts: Similar functionalities, specifically tailored for capacity-based constraints.
  • Branching Module: A state-of-the-art branching selection mechanism, supporting advanced branching strategies and “learning to branch” methodologies.
  • Deluxing Package: Aggressive column reduction that can significantly improve solver performance, especially for large-scale instances.

New in 2.0

  1. Enhanced Modular Design

    • The codebase has been reorganized for greater clarity, making it easier to integrate and extend each module independently.

  2. Improved Performance

    • Faster branching decisions and more efficient memory usage, especially when handling large-scale VRPs.
    • Updated data structures for cut generation and branching, reducing overhead and speeding up solution times.

  3. Better Documentation and User Experience

    • A reorganized “Get Started” section to guide new users through installation and initial usage.
    • Additional examples and clearer references for advanced features such as Rank-1 cuts, capacity cuts, and deluxing.

  4. Broader Solver Support

    • Continued improvements for Gurobi, CPLEX, and MindOpt backends.
    • Streamlined integration paths to incorporate new or custom solvers.

Development History

RouteOpt was initiated in 2021 by Zhengzhong (Ricky) You during his PhD at UF ISE, under the guidance of Dr. Yu Yang. Initially aimed at enabling research on learning-based branching, it has evolved into a flexible, modular framework. Early releases focused on the Capacitated Vehicle Routing Problem (CVRP), but RouteOpt has since been expanded to handle other VRP variants with greater ease and efficiency.

Motivation and Inspiration

A key inspiration for RouteOpt has been the need for accessible, open-source solutions that support advanced branching and cutting strategies—particularly for research in “Learning to Branch.” RouteOpt addresses these needs while encouraging further innovation in the broader routing and optimization community.

Commitment to Development

RouteOpt’s evolution continues with each release. Special acknowledgment goes to the creators of VRPSolver, whose foundational work influenced RouteOpt’s design. We strive to push the boundaries of VRP solvers by:

  • Expanding to more VRP variants.
  • Enhancing scalability, adaptability, and flexibility.
  • Providing more robust parameter management and stability.

These areas offer exciting opportunities for the community to collaborate and shape the future of exact VRP solving.

Getting Started

Thank You

We appreciate your interest in RouteOpt and look forward to your feedback and contributions. If you find this project useful in your research or applications, please consider citing it. We hope RouteOpt 2.0 significantly enhances your VRP-solving experience—happy optimizing!

Download
Grab the source code from this release page or clone our repository at:

git clone https://github.com/Zhengzhong-You/RouteOpt.git
Assets 2
Loading