Context:

This research solves the problem of finding the optimal location point for a fleet of garbage collection vehicles, as well as their optimal routes in order to minimize the cost of garbage collection in 144 neighborhoods of the municipality of Dosquebradas, Risaralda, Colombia, using 8 vehicles with homogeneous capacity of 25 tons which belong to the company Serviciudad.

Methods:

Firstly, a scanning heuristic (Ospina Toro, Toro Ocampo, & Orrego Cardozo, 2016) was used to find a good point of departure for the group of all the vehicles in order to generate good-quality initial routes. Then, these initial routes feed the modified genetic algorithm of Chu-Beasley (Solarte Martinez, Castillo Gaspar, & Rodriguez, 2015), taking into account the load capacity of the vehicles. Finally, in search of an optimal result, the best result found in the previous step is treated again with a Tabu metaheuristic (Bodas López, 2017).

Results:

A new methodology was designed and called hybrid CSGTR (Clustering, sweep, genetic y tabú routing), which allows to take advantage of clustering before vehicle routing (Rueda Bayona, Elles Pérez, Sánchez Cotte, González Ariza, & Rivillas Ospina, 2017) and includs Heuristic models such as Scanning Technique and Metaheuristic models like the Chu-Beasley’s and Tabu Algorithm. The application of the CSGTR methodology allowed to reduce the time and costs of the routes of garbage trucks in the municipality of Dosquebradas, Risaralda, Colombia.

Conclusions:

The hybrid methodology CSGTR, to solve the problem of location of vehicle fleets and generation of collection routes, is presented as an alternative approach with better results than the previous approach.