Urban Transportation and Logistics

204 Ali Gul Qureshi

theneighborhood of all customers, is used in order to reduce the com-

putation time for the formed local search. In the second stage, Branchini

etal. (2009) have used an adaptive granular local search in which the

cardinality of the set A′ is increased or decreased with the decrease or

increase in the arrival rate of immediate requests respectively.

Metaheuristics such as tabu search (TS) and genetic algorithms (GAs)

have been successfully used for the static VRPTW. In the DVRPTW

case, their relatively higher computational requirement poses some

problems, especially when rejection policy is adopted. Nonetheless, an

appropriate implementation can yield a much better optimization com-

pared to the local optimization- base