查看更多>>摘要：Cellular networks are one of today's most popular means of communication. This fact has made the mobile phone industry subject to a huge scientific and economic competition, where the quality of service is key. Such a quality is measured on the basis of reliability, speed and accuracy when delivering a service to a user no matter his location or behaviour are. This fact has placed the users’ tracking process among the most difficult and determining issues in cellular network design. In this paper, we present an adaptive bi-phased evolutionary algorithm based on the takeover time to solve this problem. The proposal is thoroughly assessed by tackling twenty-five real-world instances of different sizes. Twenty-eight of the state-of-the-art techniques devised to address the users’ mobility problem have been taken as the comparison basis, and several statistical tests have been also conducted. Experiments have demonstrated that our solver outperforms most of the top-ranked algorithms.
查看更多>>摘要：During the last decade, the Integrated Circuit industry has paid special attention to the security of products. Hardware-based vulnerabilities, in particular Hardware Trojans, are becoming a serious threat, pushing the research community to provide highly sophisticated techniques to detect them. Despite the considerable effort that has been invested in this area, the growing complexity of modern devices always calls for sharper detection methodologies. This paper illustrates a pre-silicon simulation-based technique to detect hardware trojans. The technique exploits well-established machine learning algorithms. The paper introduces all the background concepts and presents the methodology. The validity of the approach has been demonstrated on the AutoSoC CPU, an industrial-grade, safety-oriented, automotive benchmark suite. Experimental results demonstrate the applicability and effectiveness of the approach: the proposed technique is highly accurate in pinpointing suspicious code sections. None of the hardware trojans from the set has been left undetected.
查看更多>>摘要：This paper presents a new quantum-behaved particle swarm optimization (QPSO) algorithm for the trajectory planning task of mobile robotic vehicles in static and dynamic environments—it is called enhanced diversity particle swarm optimization (EDPSO). The main characteristic of this algorithm is that it has peaks of diversity in its population, making it possible to escape from local minima effectively, avoiding stagnation. Through the proposed PSO, it is possible to obtain safe and efficient routes, avoiding energy waste and maintaining system integrity in several possible applications. The parameters of the proposed algorithm were tuned using the benchmarking functions. The same functions were used to compare the algorithm with those already established in the literature. Once the proposed algorithm showed promising results, it was simulated in four environments, each with different complexities, presenting dangerous regions and terrains unsuitable for robot navigation, and a large number of obstacles or even moving objects. Further, the algorithms used for comparison were also simulated and the EDPSO presented satisfactory results. Through simulations it was possible to notice that the proposed approach resulted in collision-free and planned routes, and the algorithm presented increased exploration features owing to the diversity peaks that occur during the optimization.
查看更多>>摘要：This paper presents an interval multi-objective path planning (PP) scheme for patrol robot in nuclear power plant. The purpose of this PP scheme is to find collision-free paths with the shortest length and smallest risk degree. Firstly, a novel workspace modeling method is proposed to describe the static PP environment of patrol robot in nuclear power plant. Then considering the conflicts of the shortest length and smallest risk degree, an interval multi-objective particle swarm optimization (IMOPSO) method is used. In the IMOPSO, an ingenious interval update law for the particle's global best position and local best position based on the crowding distance of each risk degree interval is used to increase the diversity of population, and an iterative procedure is adopted to update the particle's position when the found paths are collided with some existing obstacles. Finally, three representative simulation tests are used to verify the validity of proposed IMOPSO method. Results show that comparing with other three well-known multi-objective evolutionary algorithms, our proposed method has the advantages of finding a better Pareto optimal paths.
查看更多>>摘要：In the current context of pursuing carbon neutrality and carbon peaking, many countries are accelerating the construction of carbon trading markets. Accurate prediction of carbon prices can enable national carbon trading markets to play a role in carbon emission reduction as soon as possible. However, current research is limited mostly to point forecasting of carbon prices, which makes it difficult to guarantee the stability of forecasting results in an increasingly complex market. Therefore, this paper proposes a three-stage vertical carbon price interval prediction framework. The contributions of this paper are as follows: the selection process of the decomposition model is regarded as an important process of prediction; a backpropagation neural network optimized by the sparrow search algorithm (SSA-BPNN) is used for the point prediction of carbon prices as a first attempt; and the kernel density estimation (KDE) model is used for interval estimation based on the point prediction results, which improves the confidence of the prediction. To validate the framework, this paper uses Shenzhen SZA-2014 products as the sample. The results show that the root mean square error of the predicted result with the improved complete ensemble empirical mode decomposition model (ICEEMD) is reduced by 29.7% and the use of SSA increases the predicted R2 by 8.5% compared with other optimization algorithms. In addition, the prediction interval coverage probability of interval prediction reaches 86% under 70% confidence. These results show that the proposed framework is not only more effective in point prediction but also performs well in interval prediction.
查看更多>>摘要：In this work, we proposed a face feature extraction method by Learning Gabor Log-Euclidean Gaussian with Whitening Principal Component Analysis (called LGLG-WPCA). The proposed method extracts raw features from the multivariate Gaussian in the transform domain of Gabor wavelet and uses WPCA to get robust features. Because the space of Gaussian is a Riemannian manifold, it is difficult to incorporate the learning mechanism into the model. To address this issue, Log-Euclidean approach is used to embed the multivariate Gaussian into the linear space, and then use WPCA to learn discriminative face features. LGLG-WPCA is good at extracting the detail features of face image. Furthermore, another outstanding advantage of LGLG is that its features can be effectively integrated with the high-level features of deep learning network for face recognition in more complex environments. We presented the feature fusing approaches for face recognition based on Self-attention Network (SAN) and achieved obvious performance improvement to the-state-of-the-art deep networks including SENet and FaceNet. Experiments show the proposed method is robust under adverse conditions such as varying poses, skin aging and uneven illumination, and it is suitable for face image under small-scale datasets in complex environments, such as network-based or video-based person searching or tracking.
查看更多>>摘要：In this study, the operating room scheduling of hospital networks with virtual alliance has been studied, which at the same time, there is a kind of cooperation and competition among the agents. The main feature in networks with the virtual alliance is the possibility of different objective functions among the agents, which has priority for agents compared to the network's overall objective. Here, by considering the conditions of emergency arrival, the time of inter-hospital transportation, and the elective patients and non-elective patients in the scheduling, an attempt has been made to bring the problem closer to real-world situations. To solve this problem, first, a mixed-integer mathematical programming model is proposed. Because of its NP-hardness, then, a multi-objective learning variable neighborhood search algorithm is designed to minimize total completion of surgeries, the cost of allocating the patient to the hospital and the surgeon, and the cost of overtime operating rooms throughout the network. Finally, the performance of the proposed algorithm is evaluated in comparison with the NSGA-II and memetic-based algorithm, which due to considering the learning mechanism along with the use of various neighborhood structures in the proposed algorithm, its results are promising. It is expected that by using the proposed algorithms in a cooperative structure, the hospitals are able to achieve optimal/near-optimal solutions in a reasonable time, in which, in addition to more economic activity, patients also benefit due to better use of resources.
查看更多>>摘要：Over the past few decades, personnel–position matching (PPM) has garnered increasing attention from scholars. In recent years, with intelligent robots facilitating intelligent production, a new form of problem—personnel–machine position matching (PMPM)—has been derived from PPM. In this study, an interval-valued intuitionistic fuzzy two-sided matching model considering level of automation (LOA) is proposed to solve the PMPM problem in an intelligent production line from the perspective of position homogeneity. The first issue to be considered in this solution is the uncertainty of preference information resulting from the decision-makers’ cognition bias or limitations. By proposing a novel score function based on the centroid method and technique for order preference by similarity to an ideal solution (TOPSIS), this issue is addressed in the information evaluation phase. Another important issue lies in the LOA, which adjusts the degree of human–machine participation. In this study, the classical two-sided matching model was improved by considering the LOA. Furthermore, to maximise the matching satisfaction of multiple sides (personnel, intelligent robots and positions), a multi-objective decision-making model is established. Afterwards, the model is transformed into a single objective model using the combined satisfaction analysis method, which is introduced to produce the final optimisation results in this modelling process. A case is presented to illustrate the practicality of the interval-valued intuitionistic fuzzy two-sided matching model considering LOA, and the results indicate that this model can solve the PMPM problem in an intelligent production line.
查看更多>>摘要：The robustness of landslide prediction models has become a major focus of researchers worldwide. We developed two novel hybrid predictive models that combine the self-organizing, deep-learning group method of data handling (GMDH) with two swarm intelligence optimization algorithms, i.e., cuckoo search algorithm (CSA) and whale optimization algorithm (WOA) for spatially explicit prediction of landslide susceptibility. Eleven landslide-causing factors and 334 historic landslides in a 31,340 km2 landslide-prone area in Iran were used to produce geospatial training and validation datasets. The GMDH model was employed to develop a basic predictive model that was then restructured and its parameters were optimized using the CSA and WOA algorithms, yielding the novel hybrid GMDH-CSA and GMDH-WOA models. The hybrid models were validated and compared to the standalone GMDH model by calculating the area under the receiver operating characteristic (AUC) curve and root mean square error (RMSE). The results demonstrated that the hybrid models overcame the computational shortcomings of the basic GMDH model and significantly improved landslide susceptibility prediction (GMDH-CSA, AUC = 0.909 and RMSE = 0.089; GMDH-WOA, AUC = 0.902 and RMSE = 0.129; standalone GMDH, AUC = 0.791 and RMSE = 0.226). Further, the hybrid models were more robust than the standalone GMDH model, showing consistently excellent performance when the training and validation datasets were changed. Overall, the swarm intelligence-optimized models, but not the standalone model, identified the best trade-offs among objectives, accuracy, and robustness.
查看更多>>摘要：We present a novel framework that integrates segmentation of lesion masks and prediction of COVID-19 in chest CT scans in one shot. In order to classify the whole input image, we introduce a type of associations among lesion mask features extracted from the scan slice that we refer to as affinities. First, we map mask features to the affinity space by training an affinity matrix. Next, we map them back into the feature space through a trainable affinity vector. Finally, this feature representation is used for the classification of the whole input scan slice. We achieve a 93.55% COVID-19 sensitivity, 96.93% common pneumonia sensitivity, 99.37% true negative rate and 97.37% F1-score on the test split of CNCB-NCOV dataset with 21192 chest CT scan slices. We also achieve a 0.4240 mean average precision on the lesion segmentation task. All source code, models and results are publicly available on https://github.com/AlexTS1980/COVID-Affinity-Model.
NSTL
Elsevier