By deploying the network functions (NFs) as software applications called Virtual Network Functions (VNF), Network Function Virtualization (NFV) helps the Service Provider (SP) to reduce the Capital Expenditure (CAPEX), Operation Expense (OPEX) and save a lot of deployment time. However, there are still many challenges that remain to be solved in NFV (e.g., Virtual Network Function Placement (VNFP) problem, Traffic Routing (TRR) problem, and NF Redeployment and Consolidation (VRC) problem). We propose two different solutions to the VNFP and TRR problems under the two different scale network topologies. First, we formulated a Mixed-Integer Linear Programming (MILP) and investigated the MILP model with a gap value (MILPWG). Then we proposed an efficient heuristic algorithm called Genetic Algorithm (GA).We have implemented two different scale tree network topologies, one is composed of four servers and ten switches, and the other is composed of eight servers and twenty switches. The simulation results show that the MILPWG can obtain a near-optimal or even accurate result in less than one second when the number of VNFs is not large or the network scale is small. But MILPWG is also inefficient when the network scale is large or there are many VNF requirements. Regarding the proposed GA, even though the computation time may increase slightly when there are more VNFs, GA can get near-optimal results within 1 second regardless of the network scale.