without exceeding the vehicle’s capacity. In practice, it is common
for the clients to specify a time window in which they want to be
visited. This gives rise to the VRP with time windows (VRPTW),
which has been tackled with both heuristic and exact methods
(Baldacci, Mingozzi, & Roberti, 2012; Yu, Yang, & Yao, 2011).
One of the VRPTW variants considers two types of customers,
termed linehaul and backhaul (VRPBTW). The backhaul customers
are served after all of the linehaul customers on the route have
been visited. Ropke and Pisinger (2006) presented an extensive
study of this problem and its variants. Freight transport planning
with VRPBTW is highly applicable for decreasing the environmental
impact of freight transport because this method involves two
services that are commonly rendered separately along the same
route, thereby increasing vehicle use. To consider environmental
and energy aspects in route planning, existing mathematical models
must be analyzed. Consequently, new computational approaches
to practical solutions for these types of problems must
be proposed.
The incorporation of themes such as the minimization of energy
and CO2 emissions in routing problems is a relatively recent topic
addressed in the literature. Kara, Kara, and Yetis (2007) proposed a
new objective function that considers minimization of the product
between the load and the distance traveled by the vehicle in the
CVRP, which they termed the Energy Minimizing Vehicle Routing
Problem. The model was tested using two examples from the literature,
and the solution differed from that of the classical VRP because
the energy also depends on vehicle load. However, the load
between two nodes is only one of the many variables that influence
the energy required by a vehicle. Bektas and Laporte (2011)