A new hybrid algorithm for solving the capacitated vehicle routing problem (CVRP)
with stochastic demands is proposed. This new approach combines the simulated
annealing (SA) with the savings algorithms and is implemented to obtain results of
several Solomon’s instances of CVRP. This approach is named simulated annealing
with metropolis cycle based on savings algorithm (SAMCSA). SA is a bio-inspired
metaheuristic, an emulation of the heating and cooling processes of solids to solve
optimization problems. On the other hand, the savings algorithm is a deterministic
heuristic for solving CVRP. In order to generate high quality solutions of CVRP, this
approach applies the savings algorithm into metropolis cycle of simulated annealing.
An initial solution of SA is also generated by the savings algorithm. This simple
change has lead to increase the solution quality for the CVRP with respect to the
classical simulated annealing and savings algorithms.