This paper proposes an adaptation of PSO (Particle Swarm
Optimization) on the studied problem HVRPMBTW (VRP
with constraints Heterogeneous Fleet, Mixed Backhauls, and
Time Windows). This method is compared with EM (Exact
Method: solver CPLEX) and ACO (Ant Colony Optimization),
so the results show that the metaheuristics PSO and
ACO are effectiveness on small problems compared with
EM. The PSO algorithm is tested on a set of 24 instances and
it is tested with different parameter values. It has found new
best solutions in 4 of 24 instances, and in 14 of 24 instances
the obtained solution are equal to the best known solution
(ACO). So, the results proves the efficiency of the proposed
algorithm. Future works could include the development of a
cooperative PSO where PSO could be hybridized with the solver CPLEX.