We implement and evaluate the performance of two methods to reduce the number of decision variables used in the search for the optimal solution to CVRP (Capacitated Vehicle Routing Problem), which is one of the combinatorial optimization problems. When reducing decision variables in consideration of the number of routes, we evaluated two methods: one is to reduce decision variables by fixing some values of decision variables and the other is to reduce the generation of decision variables themselves. The former method has a limited effect of speeding up the process by decision variable reduction. The latter method, however, achieved a speedup of approximately 2 times, although it depends on the number of cities visited.