Radio frequency (RF) energy harvesting is one of the effective methods to deal with the energy limitation of wireless network nodes. The deployment positions and transmit power setting of RF energy sources (ESs) determine the energy harvesting rate of each node. Most of the existing research work considers scenarios where no candidate ES deployment positions are given. However, in the practical application scenario, there are often many areas inside the network region where the ESs cannot be placed. The ESs can only be arranged in some reasonable candidate locations. So far, almost no work has been done to study how to select appropriate deployment positions among candidate deployment positions of ESs. Given the nodes' locations, nodes' energy energy-harvesting-rate demand, the number of ESs and the candidate deployment positions of ESs, design the ES deployment schemes which minimize the total network power consumption. Firstly, the problem is modeled as a mixed integer programming problem. Then a low-complexity approximation heuristic scheme and a genetic algorithm based deployment scheme with lower total network power consumption are proposed, respectively. Simulation results show that the proposed two schemes reduce the total network power consumption by about 90% as compared to the scheme of randomly selecting the deployment locations and the total network power consumption of genetic scheme can be 35% lower than that of heuristic algorithm. Therefore, the deployment scheme based on genetic scheme can be used for the small and medium-sized ES deployment scenarios, while the heuristic scheme can be used for large-scale ES deployment scenarios.