Through the implementation of internet of things (IoT) edge computing in different application domains, end users have the challenge of coping with a constantly increasing amount of information originating from IoT edge devices while being valuable to different stakeholders. However, due to the lack of trust between node and the IoT edge devices, such devices cannot exchange large amounts of data. Performing complex processing tasks on IoT edge devices is still a challenging task due to the low processing capacity of IoT edge devices. However, for increasing computational efficiency, a power-of-collaboration (PoA) consensus mechanism is proposed in which the devices engage in the creation of blocks through the use of PoA credit scores. Also, double deep Q-network (DDQN) algorithm is proposed to develop a new transaction offloading unit named “unit,” which ultimately decreases the storage of a blockchain. A series of tests are performed in order to show the effectiveness of the presented solution when compared to existing schemes, also attained the 91% of processing power.