Contrary to most simpler vehicle routing problems, few population
or multi-agent metaheuristics have been developed for location-routing
problems. In regards to genetic algorithms, the
reason probably resides in the necessity to design chromosome
encodings which capture the location and routing dimensions,
with ad hoc crossover operators. Multi-agent methods such as
ant colony algorithms are traditionally based on the construction
of one route by each ant, and implementations become less obvious
when location decisions are added