TITLE:
Experiment Design for the Location-Allocation Problem
AUTHORS:
María Beatríz Bernábe Loranca, Rogelio González Velázquez, Martín Estrada Analco, Mario Bustillo Díaz, Gerardo Martínez Guzman, Abraham Sánchez López
KEYWORDS:
Demand, Experimental Design, Heuristics, Location-Allocation, Partitioning.
JOURNAL NAME:
Applied Mathematics,
Vol.5 No.14,
July
31,
2014
ABSTRACT:
The
allocation of facilities and customers is a key problem in the design of supply
chains of companies. In this paper, this issue is approached by partitioning
the territory in areas where the distribution points are allocated. The demand
is modelled through a set of continuous functions based on the population
density of the geographic units of the territory. Because the partitioning
problem is NP hard, it is necessary to use heuristic methods to obtain reliable
solutions in terms of quality and response time. The Neighborhood Variable
Search and Simulated Annealing heuristics have been selected for the study
because of their proven efficiency in difficult combinatorial optimization
problems. The execution time is the variable chosen for a factorial
experimental design to determine the best-performing heuristics in the problem.
In order to compare the quality of the solutions in the territorial partition,
we have chosen the execution time as the common parameter to compare the two
heuristics. At this point, we have developed a factorial statistical
experimental design to select the best heuristic approaches to this problem.
Thus, we generate a territorial partition with the best performing heuristics
for this problem and proceed to the application of the location-allocation
model, where the demand is modelled by a set of continuous functions based on
the population density of the geographical units of the territory.