SD-Jaya Based Multi-Objective Optimization Algorithm for IRS-Aided Air-to-Ground Task Offloading in Charging Electric Vehicle Networks

The integration of electric vehicles (EVs) into the power grid has led to a significant increase in load demand. However, the absence of efficient edge computing services for charging EVs poses challenges to the grid's safe and stable operation. Our study proposes an intelligent reflective surface (IRS)-assisted air-to-ground task offloading network for charging EVs that utilizes tethered unmanned aerial vehicles (tUAVs) equipped with IRS for task transmission to provide efficient edge computing services. The network architecture includes single-IRS and double-IRS links to enhance communication efficiency, while EVs are equipped with Network in Box (NIB) units to provide flexible computing power. We formulate a multi-objective optimization problem aimed at minimizing transmission obstacle, power purchase cost, transmission delay, and failure rate in the network. Since the optimization problem is NP-hard, we propose an improved Self-Learning Discrete Jaya (SD-Jaya) algorithm which finds the offloading strategy using historical knowledge and the selection of strategy-related operators. Moreover, we develop a novel tournament-based algorithm to select the Pareto layer in the multi-objective problem. Simulation results show that the proposed algorithm is able to find better non-dominated solutions. The power purchase cost of the proposed scheme is 43% and 54% lower than that of the single-IRS scheme and the local scheme, respectively.