Unused channel sensing and allocation in Cognitive Radio Networks (CRN) are streamlined using guard intervals to enhance the Quality of Service (QoS) of secondary users (SUs). Identifying the guard interval succeeding unused channels is tedious due to multi-interference and asynchronous resource allocations. To address this problem, the article introduces a Multi-Objective Optimization Scheme (MOOS) using Partial-Order Transfer Learning (POTL). The objectives are defined as interference minimization, optimal power utilization, and maximum channel allocation. These objectives are satisfied by untying the allocated channels that are close and away from the guard interval. The guard interval without overlapping channels is identified to reduce the allocation failures by mitigating the asynchronous intervals. The partial order derivatives are responsible for verifying the asynchronous overlap between the guard and unused channel allocation intervals. From this estimation, the overlapping and non-overlapping intervals are distinguished to improve the allocation with optimal power utilization. The transfer learning opts for the selection of a precise unused channel that satisfies maximum QoS objectives during SU response intervals. The proposed optimization scheme improves the channel allocation rate by 10.13%, power utilization efficiency by 10.5%, and network throughput by 12.03% for the maximum number of secondary user variants considered.
K. et al. (Sun,) studied this question.