Neighbourhood functions are key components of local-search algorithms such as Simulated Annealing or Threshold Accepting. These functions take a solution and return a slightly-modified copy of it, i.e. a neighbour. The package provides a single function, neighbourfun(), that produce a neighbourhood function, based on various parameters such as minimum or maximum values. Currently supported are numeric and logical solutions. The algorithms were originally created for portfolio-optimization models, but can also be used for other models.
The latest released version is available from http://enricoschumann.net. In an R session, just type:
install.packages('neighbours,
repos = c('http://enricoschumann.net/R', getOption('repos')))