| dc.description.abstract | The current research mentions the ineffective delivery rate of Viettel Post recently due to a
prominent factor of long shipment duration or high distribution’s travel time which is possibly
defined by the large distance between hub locations and assigned post offices. Thus, a hub
location problem is established to further investigate the optimal locations in the hub potential
set along with the non-hub nodes’ multi-allocation and their corresponding routing between
those assigned non-hubs and hubs. The multi-allocation mechanism allows the non-hub nodes
to be allocated to at least one hub which is much practical in real scenarios. The expected
outcome considers the number of hubs to be opened, the locations and equivalent allocations,
and vehicle routing to the problem with the goal of operation network costs minimization.
A Mixed Integer Linear Programming (MILP) is proposed to obtain the exact solutions and a
metaheuristic – Adaptive Large Neighborhood Decomposition Search (ALNDS) is then
applied for validation whether which method gives more high-quality performance.
Furthermore, the six datasets with small, medium, and large-sized instances are generated to
better evaluate the logic and accuracy of the mathematical model as well as compare the
corresponding sizes with the algorithm. The results present the ALNDS yields more optimal
solutions under shorter computational time for the medium to large-sized instances and
slightly equal to or bigger than the exact solutions in small-sized instances. In addition, the
sensitivity analysis on both the unit transportation costs and vehicle capacity along with the
fleet size delivers higher costs when increasing those parameters and vice versa. | en_US |
