Eiselt [19] evaluates how far the observed locations of landfills
and transfer stations deviate from the optimized (i.e., cost-minimizing)
solution. The location problem is modeled through a pathbased
garbage transfer problem, in which the decision to locate
landfills and transfer stations at some sites is made by means of
binary variables. He develops a 1/S,L/–/TC MIP model, in which the
proportion of the garbage shipped directly from the customers to
the landfills and via transfer stations is represented by flow
allocation variables. Similar to hub location problems, a discount
is charged into the objective function whenever the garbage is
collected at the transfer stations and delivered to the landfills,
rather than shipped from the customers to the transfer stations or
landfills. Computational experiments are carried out with the aim
of comparing the performance of the proposed model with the
location plan implemented for the province of New Brunswick,
located in Eastern Canada