The increasing interest in distributed energy generation from renewable sources is enabling traditional energy consumers to become active energy producers. They can be formed into virtual clusters for easier management and to reduce costs; the virtual clusters are usually referred to as virtual microgrids (VMG). The VMGs are coordinated by energy trading agents (ETA), a communication hardware or software, Which coordinates a population of prosumers of a certain size. We examine the case when prosumers communicate within VMGs via low-power wide area network (LP-WAN) technologies, such as LoRaWAN, whose spreading factor (SF) property affects the coverage distance and, consequently, the size of the served population of prosumers. The SF property enables the transformation of VMGs into dynamic schemes (i.e. varying number of prosumers are seen per trading period). Based on this formulation, we propose two energy trading cost models: one for the energy and one for the LoRaWAN communication system. Results show that the optimal techniques prescribed in this study can reduce energy trading cost by 52% and energy consumption for the LoRaWAN system by up to 45%. Lastly, we formulate a closed form relationship to demonstrate that bit energy decays with increasing distance for varying SF values.