Evolutionary multitasking algorithms use information exchange among individuals in a population to solve multiple optimization problems simultaneously. Negative transfer is a critical factor that affects the performance of evolutionary multitasking algorithms. In this study, we propose an innovative approach to mitigate negative transfer in evolutionary multitasking algorithms. The proposed approach is grounded in rigorous theoretical analysis, which provides valuable theoretical insights into the design of an optimal linear crossover operator for mitigating negative transfer. By identifying interpretable conditions, we establish a solid theoretical foundation to prevent negative transfer in diverse scenarios. Building upon these findings, we theoretically derive a closed-form expression for the optimal crossover operator and propose practical design methods based on approximations. Furthermore, we integrate the proposed optimal crossover operator into a fundamental evolutionary multitasking algorithm framework. The resultant algorithm is comparable or superior to other state-of-the-art methods. Empirical validation through comprehensive experiments confirms the effectiveness of our theoretical findings.

Mobile-edge computation edge servers (ESs) act as crucial intermediaries between mobile devices (MDs) and cloud computing centers, ensuring low-latency computation and reliable IoT network performance. However, regional failures induced by natural or man-made disasters can severely disrupt ES operations. This paper focuses on devising an effective repair scheme for ESs after large-scale disasters, particularly under the uncertainty of secondary disasters and the stochastic demands of MDs. We formulate the edge server repair problem (ESRP) as a mixed-integer linear programming (MILP) model and prove its NP-hardness. To effectively solve this problem, we propose a solution algorithm based on Benders decomposition (BD). The MILP problem is decomposed into a more solvable master problem and a series of subproblems, which are solved alternately and iteratively. Furthermore, we enhance our BD algorithm by incorporating two types of initial cuts to reduce the iterative number and accelerate the convergence speed. The first type is known as supermodular cut, which exploits the inherent properties of the problem. The second type is called aggregation capacity constraint, which exploits the capacity characteristics of ESs. Extensive simulation results demonstrate that the proposed method excels in ES repairing under large-scale IoT network failure scenarios. Compared to other available algorithms, our BD-based approach more efficiently achieves the exact solution to the ESRP, minimizing network demand loss within limited resources.

Vision-based vehicle lateral localization has been extensively studied in the literature. However, it faces great challenges when dealing with occlusion situations where the road is frequently occluded by moving/static objects. To address the occlusion problem, we propose a highly robust lateral localization framework called multilevel robust network (MLRN) in this article. MLRN utilizes three deep neural networks (DNNs) to reduce the impact of occluding objects on localization performance from the object, feature, and decision levels, respectively, which shows strong robustness to varying degrees of road occlusion. At the object level, an attention-guided network (AGNet) is designed to achieve accurate road detection by paying more attention to the interested road area. Then, at the feature level, a lateral-connection fully convolutional denoising autoencoder (LC-FCDAE) is proposed to learn robust location features from the road area. Finally, at the decision level, a long short-term memory (LSTM) network is used to enhance the prediction accuracy of lateral position by establishing the temporal correlations of positioning decisions. Experimental results validate the effectiveness of the proposed framework in improving the reliability and accuracy of vehicle lateral localization.

Dataset bias in stance detection tasks allows models to achieve superior performance without using targets. Most existing debiasing methods are task-agnostic, which fail to utilize task knowledge to better discriminate between genuine and bias features. Motivated by how humans tackle stance detection tasks, we propose to incorporate the stance reasoning process as task knowledge to assist in learning genuine features and reducing reliance on bias features. The full stance reasoning process usually involves identifying the span of the mentioned target and corresponding opinion expressions, such fine-grained annotations are hard and expensive to obtain. To alleviate this, we simplify the stance reasoning process to relax the granularity of annotations from token-level to sentence-level, where labels for sub-tasks can be easily inferred from existing resources. We further implement those sub-tasks by maximizing mutual information between the texts and the opinioned targets. To evaluate whether stance detection models truly understand the task from various aspects, we collect and construct a series of new test sets. Our proposed model achieves better performance than previous task-agnostic debiasing methods on most of those new test sets while maintaining comparable performances to existing stance detection models.

We propose two new algorithms called BiLAD and ExactBiLAD for the well-known Single-Constrained Shortest Path (SCSP) problem. It is a fundamental problem in quality-of-service (QoS) routing, where one seeks a source-destination path with the least cost satisfying a delay QoS constraint in a network. As pointed out by Jüttner et al., there is no widely accepted algorithm with polynomial time to the SCSP problem because the SCSP problem is NP-hard. The remarkable feature of BiLAD is that it ensures that the length of iteratively updated angle interval is shrunk at least at a constant ratio. With the help of this feature, we prove its polynomial time complexity. To the best of our knowledge, this is the first time that the polynomial time complexity is proved in details. The numerical results show that, in most QoS routing test instances, the performance of BiLAD is close to their primal optimal solutions. The proposed modified Dijkstra procedure, whose complexity is the same as that of the Dijkstra algorithm, also accelerates BiLAD. In the second part of the paper, based on the information obtained by BiLAD, we design an exact algorithm-ExactBiLAD, in which an optimal solution to the SCSP problem is finally obtained by scanning the steadily reduced optimal-path-candidate triangle area. The simulation results indicate that ExactBiLAD needs only a dozen times of executing the modified Dijkstra algorithm regardless of the network size or the average node degree. Distinguished from many other exact algorithms, ExactBiLAD has a satisfactory performance in the practical computation.

Detecting stance from certain types of question-answer pairs is an interesting problem which has not been carefully explored. Unlike previous stance detection tasks, targets here are not given entities or claims but entire questions, which makes it difficult to capture the semantics of targets and build target-dependent representations of answers. To address them, we introduce the Recurrent Conditional Attention (RCA) model which incorporates a conditional attention structure into the recurrent reading process. RCA iteratively guides the distillation of question semantic with answer information and collects stance-oriented text relating to question, further revealing mutual relationship among stance, answer and question. Experiments on a manually labeled Chinese community QA stance dataset show that RCA outperforms four strong baselines by average 2.90% on macro-F1 and 2.66% on micro-F1 respectively.

UHVDC hierarchical connection to AC grid is a specially-designed mode in China to enhance the support from the receiving AC grid to HVDC. Besides the merits to optimize power flow, such a connection mode features stronger stability than the mode in which all power is fed into the grid with a single voltage grade. In this paper, the stability of UHVDC hierarchical connection is quantitatively discussed based on the proposed mathematical model for small signal analysis, where the interactions between the converters through both AC and DC system, as well as the effects of HVDC control are taken into account. Meanwhile the sensitivity of stability to the supporting strength of AC system and the parameters of control system is demonstrated based on simulation cases built in PSCAD. Hopefully the results can facilitate the proper decision on the control parameters and the investigation on the advantage of UHVDC hierarchical connection.