POMDP Formulation for Energy Harvesting Network with Wireless Distributed Computing

Energy harvesting network (EHN) is a trending topic in wireless communications that enables wireless devices to harvest energy from ambient sources such as RF waves. However, EHN faces energy causality constraint that turns the energy harvester into sleep mode. This, in turn, delays the processing of next task until harvesting the required energy for processing. The objective of this research is to prolong the active mode of energy harvesters by employing a novel scheme of EHN that combines local computing and wireless distributed computing in one EHN. Accordingly, a novel EHN algorithm that employs partially observable Markov decision process (POMDP) is proposed in order to automate the decision of the energy harvester based on the channel variations and stored energy. Herein, the scheme of the proposed EHN is explained first. Then, the problem of the proposed EHN is formulated as POMDP. Furthermore, a policy is found in order to estimate the validity of the proposed EHN against the conventional EHN.