Latency Optimization for IRS-assisted Wireless Powered Mobile Edge Computing in loT System

The integration of intelligent reflecting surfaces (IRS) into wireless power transfer (WPT) systems in conjunction with mobile edge computing (MEC) has recently gained attention for performance enhancement. However, the latency optimization problem has not been sufficiently studied in this new scenario. In this paper, to address latency bottlenecks in distributed computing, we implement IRS-enhanced WP-MEC architecture that significantly reduces end-to-end task execution time through intelligent radio environment reconfiguration. We develop the optimization problem by jointly optimizing the uplink and downlink IRS reflection coefficients, the ratio of offloading tasks and the computing frequency. The inherent non-convexity arising from variable coupling is mitigated through a decomposition-based approach. To deal with the min-max problem in the resourse allocation optimization, A bisection search algorithm is developed to efficiently locate the optimal solution within the feasible region. To address the non-convexity of the rank-one constraint, a penalty-based approach is proposed for efficient phase shift design in wireless energy transfer (WET). Simulation results confirm that integrating IRS into the WP-MEC system yields substantial performance improvements compared to benchmark methods.