Exact Noise-Robust Distributed Gradient-Tracking Algorithm for Constraint-Coupled Resource Allocation Problems

The Industrial Internet of Things (1IoT) gradually becomes a new paradigm for information exchange in the industrial production environment. To ensure the high reliability of IloT services, an efficient resource allocation method with good robustness is urgently needed under complex industrial environments. This paper considers the distributed constraint- coupled resource allocation problem with noisy information exchange over an undirected network, where each agent holds a private cost function and obtains the solution via only local communications. Communication noise poses a challenge to gradient-tracking based algorithm as the impact of noise will accumulate and its variance tends to infinity when the noise is persistent. Adopting noise-tracing scheme, we propose an exact noise-robust distributed gradient-tracking algorithm to achieve cost-optimal distribution of resources, which can avoid noise-accumulation in the tracking step. Moreover, noise suppression parameters are introduced to further attenuate the impact of noise. With diminishing suppression parameters, it is theoretically proved that the proposed algorithm is able to achieve exact convergence to the optimal solution. Finally, a numerical example is provided for verification.