Internet of Things (IoT) as an emerging technology is going to play a critical rule in the physical-cyber ecosystem. In this environment, sensing and actuating are considered as services for building intelligent applications. From performance point of view, a main question is how to locate these services on devices can realize a remarkable cost-efficiency and energy saving. In this paper, the goal is to specify which accompanying service with another specific service can be set on a device to optimize the energy consumption, load amount on devices, and the service response time. For this purpose, we formulate this goal as a multiobjective decision-making problem, and then we solve it by using the PROMETHEE method. In this solution, the utility of each device is determined and estimated based on entropy technique that used for computing weights for different criteria. Numerical simulation results show that the proposed co-locating algorithm, called p-co-locating, outperforms none-co-locating approach in terms of energy consumption, load amount, and the service response time.