TSP - Traveling Salesperson Problem - R package

This package provides the basic infrastructure and some algorithms for the traveling salesman problems (symmetric, asymmetric and Euclidean TSPs). The package provides some simple algorithms and an interface to the Concorde TSP solver and its implementation of the Chained-Lin-Kernighan heuristic.

Installation

Stable CRAN version: install from within R with

install.packages("TSP")

Current development version: Download package from AppVeyor or install from GitHub (needs devtools).

library("devtools")
install_github("mhahsler/TSP")

Usage

Load a data set with 312 cities (USA and Canada) and create a TSP object.

R> library("TSP")
R> data("USCA312")
 
R> tsp <- TSP(USCA312)
R> tsp

object of class 'TSP'
312 cities (distance   'euclidean')

Find a tour using the default heuristic.

tour <- solve_TSP(tsp)
tour

object of class 'TOUR' 
result of method 'arbitrary_insertion+two_opt' for 312 cities
tour length: 40621

An online example application of TSP can be found on shinyapps.

References

• Michael Hahsler and Kurt Hornik, TSP - Infrastructure for the Traveling Salesperson Problem, Journal of Statistical Software, 22(2), 2007.
• TSP package vignette with complete examples.
• Reference manual
• Concorde TSP solver home page