Releases: vidalt/HGS-CVRP
v2.0.0
This v2.0.0 release includes C interfaces developed by @chkwon as well as additional refactoring by @vidalt.
Upgrade Steps
The code is now compiled through a CMake and the compiler should support C++17 syntax (see README.md)
The '-bks' option in the command line is now deprecated. The rest of the options to run the code on a CVRPLIB instance file through the command line remain the same. Additional options are provided to directly set the parameters and verbosity level in the command line.
New Features and Improvements
We provide a simple C interface through the solve_cvrp(...) function. The solve_cvrp_dist_mtx(...) function also permits passing directly a distance matrix to the algorithm (instead of 2D Euclidean coordinates).
Moreover, Python and Julia interfaces for this code are currently being developed in the PyHygese and Hygese.jl projects.
The move evaluation procedures have been upgraded to follow a layered strategy, permitting to save computational time when handling additional constraints.
Contributors
Assets 2
v1.0.0
This is the original version of the HGS-CVRP algorithm, as published in [Vidal, T. (2022). Hybrid genetic search for the CVRP: Open-source implementation and SWAP* neighborhood. Computers and Operations Research, 140, 105643]. It also includes a minor fix done in this commit.