In this paper, we consider multi- crossducking vehicle routing problem (MCVRP) and close–open mixed vehicle routing problem (COMVRP) in the stochastic environment. We supposed that the fleet of vehicles is heterogeneous. Minimizing the total cost of serving customers is the objective of the prob- lem. We assume that the travel time between retailers nodes are nondeterministic. Firstly the new mathematical model is proposed and then we develop a hybrid solution methodology encompasses ro- bust optimization, firefly and genetic algorithm. Computational results demonstrate that the proposed algorithm can efficiently be used to solve the proposed model.
