Imbalanced datasets are a major challenge in industrial fault diagnosis because the majority of data belong to the non-fault class, while fault instances constitute a small minority. Genetic Programming (GP) has shown great potential in handling imbalanced classification tasks due to its ability to evolve classifiers and automatically optimize decision rules. Traditional GP methods for imbalanced classification often rely on fixed decision thresholds (e.g., 0 or 0.5). However, such thresholds fail to adapt to varying data distributions, resulting in limited accuracy in fault diagnosis. While threshold-free methods, such as GP using the Area Under the Curve (AUC) and its variants as fitness functions, have demonstrated effectiveness, practical applications in industrial systems often require explicit thresholds to generate accurate class labels. This paper introduces a GP-based approach with a simplified AUC variant as the fitness function and a dynamic threshold search mechanism. By adaptively optimizing thresholds during evolution, the method improves minority class detection. Experiments on public fault diagnosis datasets with varying imbalance ratios demonstrate that the proposed approach consistently outperforms traditional GP methods.

RF energy harvesting is a promising potential solution for providing convenient and perpetual energy supply to low-power wireless sensor networks. In this paper, we investigate the performance of overlaid wireless sensor transmission powered by RF energy harvesting from existing wireless system. Specifically, we derive the exact closed-form expression for the distribution function of harvested energy over a certain number of channel coherence time over Rayleigh fading channels with the consideration of hardware limitation, such as energy harvesting sensitivity and harvesting efficiency. We also obtain the exact distribution of the number of coherence time needed for fully charging the sensor. Based on these analytical results, we analyze the average packet delay and packet loss probability of sensor transmission subject to interference from existing system, for delay insensitive traffics and delay sensitive traffics, respectively. Finally, we investigate the optimal design of energy storage capacity of the sensor nodes to minimize the average packet transmission delay for delay insensitive traffics with two candidate transmission strategies.

This paper studies the performance of a RUB-based cognitive radio (CR) network, which coexists with a primary network with single primary transmitter (PT) and single primary user (PU). It is assumed that the PU can control the availability of each beam of the secondary base station (SBS) over a feedback link. The PU only needs to feed back to the SBS the index of the usable beams, which lead to the received SINR at the PU larger than a predefined beam selection threshold, and the outage probability of the primary system can be limited to a tolerable level. The SBS then selects one from the usable beams to achieve the largest throughput of the secondary system. We derive the accurate upper bound of the outage probability of the primary system and the exact throughput of the secondary system. Numerical examples show that to avoid the burst increase of the outage probability of the primary system, as well as achieve the maximal throughput of the secondary system, the beam selection threshold should be equal to the outage threshold of the primary system.