查看更多>>摘要：Carbon price forecasting is an important component of a sound carbon price market mechanism. The accurate prediction of carbon prices is an active topic of research. However, many previous studies have focused on the application of a single model, ignoring the application of combination strategies. In this study, a hybrid forecasting system that includes error correction strategy and divide-conquer strategy is designed to predict the carbon price series accurately. Specifically, the main framework of this article comprises four modules. Data preprocessing module of the divide and conquer strategy is proposed. Next, the optimization module uses a multi-objective grasshopper optimization algorithm to enhance the performance of the prediction module. Then, the error correction module predicts the error sequence and corrects the model results. To verify the performance of the established hybrid forecasting system, experiments were performed using two real carbon price series from China and European Union emissions trading schemes, and the results showed that the mean absolute percentage errors of the system were 2.7793% and 0.6720%, respectively, which are better than the other benchmark methods considered. Moreover, it was proved that the designed forecasting system provides a new, effective, and feasible solution for carbon price forecasting.
查看更多>>摘要：The authors regret ‘The common fund support in Acknowledgments should be the National Natural Science Foundation of China, China under grant No. U1733119, Tianjin Postgraduate Research and Innovation Project Funding, China under grant 2020YJS017, and the Scientific Research Project of Tianjin Education Commission, China under Grant 2020KJ013’. The authors would like to apologise for any inconvenience caused. We are very sorry for missing an important fund support that should have been together and apologize for the inconvenience caused.
查看更多>>摘要：Anomaly-based impersonation detection consists in constructing typical profiles based on users’ frequent behaviors and comparing them with new data. The underlying idea is that a very different behavior may indicate possible fraud, i.e., someone trying to impersonate the user. Most research in the area aims to use spatiotemporal data broadly available from ubiquitous location sensors, e.g., GPS, mobile telephony, beacons, and Physical Access Control Systems. Many studies achieved good performance in finding social bonds among users. Our previous work Silva and Sichman (2019) combined concepts from previous research and proposed using social groups to construct mobility profiles to enhance anomaly-detection. This paper extends our previous work and explores the feasibility of using spatiotemporal mobility profiles enriched with group patterns for fraud detection in Physical Access Control Systems. An empirical analysis is conducted using data from two real-world datasets, and results show that it is feasible to add companions activities information to mobility profiles to enhance anomaly-based impersonation attack detection.
查看更多>>摘要：With the increase in sophistication and connectedness of the healthcare networks, their attack surfaces and vulnerabilities increase significantly. Malicious agents threaten patients’ health and life by stealing or altering data as it flows among the multiple domains of healthcare networks. The problem is likely to exacerbate with the increasing use of IoT devices, edge, and core clouds in the next generation healthcare networks. Presented in this paper is MUSE, a system of deep hierarchical stacked neural networks for timely and accurate detection of malicious activity that leads to alteration of meta-information or payload of the dataflow between the IoT gateway, edge and core clouds. Smaller models at the edge clouds take substantially less time to train as compared to the large models in the core cloud. To improve the speed of training and accuracy of detection of large core cloud models, the MUSE system uses a novel method of merging and aggregating layers of trained edge cloud models to construct a partly pre-trained core cloud model. As a result, the model in the core cloud takes substantially smaller number of epochs (6 to 8) and, consequently, less time, compared to those in the edge clouds, training of which take 35 to 40 epochs to converge. With the help of extensive evaluations, it is shown that with the MUSE system, large, merged models can be trained in significantly less time than the unmerged models that are created independently in the core cloud. Through several runs it is seen that the merged models give on an average 26.2% reduction in training times. From the experimental evaluation we demonstrate that along with fast training speeds the merged MUSE model gives high training and test accuracies, ranging from 95% to 100%, in detection of unknown attacks on dataflows. The merged model thus generalizes very well on the test data. This is a marked improvement when compared with the accuracy given by un-merged model as well as accuracy reported by other researchers with newer datasets.
查看更多>>摘要：A hybrid heuristic is proposed to solve a location-routing problem with two-dimensional loading constraints. This problem appears in military situations and natural disasters in such a way that decisions are taken in a short time horizon. The proposed heuristic combines the simulated annealing method and the artificial algae algorithm. Simulated annealing is used to handle the location-routing problem, while the artificial algae algorithm is used to determine the sequence in which the items will be packed. Therefore, we apply the Skyline technique to find a feasible packing of such items onto the vehicle's rectangular surface. As there is no other work in the literature handling the location-routing problem with two-dimensional loading constraints that we can compare the results, we evaluate the heuristic performance on its subproblems: the location-routing problem and the vehicle routing problem with two-dimensional loading constraints. Although the heuristic is not designed for these subproblems, it still obtains competitive results, with an average relative difference of 1.26% and equal or better solutions for more than 90 instances. Regarding the problem under study, the heuristic obtains solutions close to an estimated lower bound for instances having more items per customer.
查看更多>>摘要：With the increasing product variety, companies aim to become better than their competitors by providing a superior product developed with a customer-oriented product design approach and a quality strategy. In order to achieve this, companies should well understand customer expectations and quickly be able to convert these expectations to technical characteristics. Since the expectations consist of mostly subjective judgments, this evaluation process contains vagueness and impreciseness. A triplet represents the uncertainty in subjective judgments: the degrees of belongingness or Truthiness (T), non-belongingness or Falsity (F), and indeterminacy (I). For this reason, in this paper, a neutrosophic Quality Function Deployment (QFD) methodology based on neutrosophic AHP and neutrosophic DEMATEL is developed and applied to the design of a car seat. In this methodology, the weighting of customer requirements is performed by neutrosophic AHP. The relationships among the technical characteristics are determined by neutrosophic DEMATEL for the customer-oriented product design, considering both impreciseness in the data and indeterminacy of the decision-makers. In other words, the contribution of this paper is that the proposed methodology provides better integration of the voice of customers into technical characteristics through a practical fuzzy multi-criteria decision analysis. Based on the results, it is revealed that seat height is the most important technical characteristic, followed by vertical travel range and horizontal travel range. Moreover, validity and verification of the proposed methodology have been tested with other methods presented in the literature. Sensitivity analyses have been carried out to show the flexibility of the given decisions under different cases. Lastly, possible implications on theoretical and managerial aspects have been discussed.
查看更多>>摘要：A fuzzy binary relation (for short, fuzzy relation) is a fundamental notion in fuzzy set theory. This paper proposes novel entropy measure for a fuzzy relation and considers its application to attribute reduction for heterogeneous data. We first define a new fuzzy entropy to compute the uncertainty of a fuzzy relation and then put forward the notions of joint information entropy, conditional information entropy and mutual information entropy in an information system with heterogeneous data. The proposed measure can overcome the weakness of the existing measure. Next, we apply the proposed measure to perform attribute reduction in this kind of information systems. Finally, we make experimental analysis to check the feasibility and efficiency of the proposed attribute reduction algorithms.
查看更多>>摘要：There have been growing interests in the energy-efficient production scheduling recently because of the growing shortage of energy. Open shop scheduling problem (OSSP) is a kind of common but seldom concerned production scheduling problem. This paper focuses on energy-efficient OSSP (EOSSP), where the effect of setup operation on production efficiency and energy consumption is considered. A multi-objective energy-efficient model based on machine speed scaling mechanism is proposed. To handle this multi-objective problem, we propose an effective adaptive multi-objective differential evolution (AMODE) algorithm. The AMODE uses a new fitness evaluation mechanism (FEM) based on dynamic reference point and fuzzy correlation entropy analysis to assess the solutions in evolution population. It also uses an adaptive opposition-based learning (AOBL) to improve its local search ability. Taguchi method is utilized to obtain the best combination of critical parameters of the AMODE. The proposed mathematical model is validated with CPLEX, and a lexicographic method is used to determine the preferable solution. Experimental results show that both our proposed FEM and AOBL can improve the performance of AMODE. Extensive experiments reveal that the performance of AMODE is superior to the other three well-known algorithms in addressing the EOSSP.
查看更多>>摘要：Online users are typically involved in multiple online social networks simultaneously to enjoy a variety of social network services, thus bringing about the interconnection of online social networks via bridge users called anchor users. Anchor users can be beneficial to a wide range of social network analysis applications such as cross-domain system recommendation, cross-network information diffusion, and link prediction, taking anchor user's intra-network structural information along with its cross-network structural properties into consideration. Several studies have so far tried to learn low-dimensional representations of social users by capturing their network structures inside one social network but they have not fully leveraged their intra-network structures with their cross-network structures to boost the performance of the aforementioned analysis tasks. In this paper, we present a novel deep learning model to learn Overall low-dimensional Vector Representations for Anchor Users (OVRAU), from a multiplex heterogeneous social network by investigating the intra-network as well as the cross-network structural information. Unlike previous works, our proposed model considers the multi-network scenario to encode diverse network structures of anchor users. We propose two types of embeddings to capture the different structural information of an anchor user from multiple social networks: a high-dimensional base embedding and a low-dimensional social edge embedding for each social network. In particular, we learn a function that generates social edge embeddings by sampling and aggregating structural features from an anchor user's neighborhood inside different social networks through one of three candidate aggregator functions namely mean, max-pooling and LSTM, with a self-attention mechanism. Link prediction is used as a downstream task to evaluate the effectiveness of the learned embeddings. Experiments were conducted on real-world social networks dataset, and the results demonstrate that our proposed model involving all the three variants can significantly outperform the existing network representation learning approaches when applied on the link prediction task and also achieve better performance over all compared baselines.
NSTL
Elsevier