Column Generation (CG) technique is popularly applied in solving the crew scheduling problem of large size, which is generally modeled as an Integer Linear Programming (ILP) problem. The traditional CG algorithms for ...
详细信息
In this paper, a rectangular microwave filter with rectangular groove is designed. The filter adopts symmetrical three-stage structure. The first section is the slot line waveguide section to realize the input / outpu...
详细信息
This paper introduces a broadband microwave bandpass filter. The structure of the filter is a filter cavity formed by two balanced dielectric sheets. On the two dielectric sheets, the relative face of the microwave in...
详细信息
This paper explores the application of the Bowyer-Watson algorithm for constructing Delaunay triangulations on Riemannian manifolds, with a particular focus on karst terrain and channel detection scenarios. We define ...
详细信息
This paper introduces a microwave filter with pentagram grooves, which belongs to an artificial surface plasmon (SSPPs) type microwave bandpass *** filter adopts a two-stage structure. The first section is a slot-line...
详细信息
Reentry trajectory optimization is a multi-constraints optimal control problem which is hard to solve. To tackle it, we proposed a new algorithm named CDEN(Constrained Differential Evolution Newton-Raphson Algorithm) ...
详细信息
Reentry trajectory optimization is a multi-constraints optimal control problem which is hard to solve. To tackle it, we proposed a new algorithm named CDEN(Constrained Differential Evolution Newton-Raphson Algorithm) based on Differential Evolution(DE) and *** transform the infinite dimensional optimal control problem to parameter optimization which is finite dimensional by discretize control parameter. In order to simplify the problem, we figure out the control parameter's scope by process constraints. To handle constraints, we proposed a parameterless constraints handle process. Through comprehensive analyze the problem, we use a new algorithm integrated by DE and Newton-Raphson to solve it. It is validated by a reentry vehicle X-33, simulation results indicated that the algorithm is effective and robust.
This paper explores black-box domain adaptation (BBDA) for cross-domain on-device machinery fault diagnosis. Specifically, a pre-trained black-box source model is deployed on a cloud platform with only its input-outpu...
This paper explores black-box domain adaptation (BBDA) for cross-domain on-device machinery fault diagnosis. Specifically, a pre-trained black-box source model is deployed on a cloud platform with only its input-output API accessible, and a randomly initialized target model is trained locally utilizing unlabeled target data on an edge device. In BBDA, neither raw data nor model parameters are transmitted across domains, and the only available supervised information for adaptation is API-queried predictions on unlabeled target data. However, API-queried predictions are inevitably noisy and class-imbalanced due to cross-domain distribution discrepancy and inherent category bias. Directly utilizing them for on-device training easily leads to confirmation bias, thereby degrading performance. This paper proposes a h ierarchical d ebiased s elf- s upervised l earning framework for BBDA from a novel research perspective of concurrently alleviating both confirmation and category biases. Specifically, debiased knowledge distillation is proposed to gradually enhance discriminability of the randomly initialized target model by providing debiased training predictions. An adaptive category-unbiased sample division strategy is devised to divide unlabeled target data into certain-aware and uncertain-aware sets, enabling fully exploiting intrinsic data structures and utilizing underlying sample relationships within the target domain. On this basis, debiased prototypical self-training is proposed for the certain-aware set to obtain pseudo-labels for self-training, leveraging global semantic structure. Hierarchical neighborhood adaptation is customized for the uncertain-aware set to enforce the cluster assumption for noise refinement, exploiting local clustering structure. Extensive experiments and analyses on public and practical datasets validate the proposed framework’s superiority in cross-domain generalization performance and practicality in on-device training efficiency.
This paper presents a sliding-mode-based diagonal recurrent cerebellar model articulation controller (SDRCMAC) for multiple-input-multiple-output (MIMO) uncertain nonlinear systems. Sliding mode technology is used to ...
This paper presents a sliding-mode-based diagonal recurrent cerebellar model articulation controller (SDRCMAC) for multiple-input-multiple-output (MIMO) uncertain nonlinear systems. Sliding mode technology is used to reduce the dimension of the control system. Two learning stages are adopted to train the SDRCMAC and to improve the stability of the control system. Lyapunov stability theorem and Barbalat's lemma are adopted to guarantee the asymptotical stability of the system. Performance is illustrated on a two-link robotic control and motor control of the human arm in the sagittal plane.
The traditional experimental teaching method is single, and students lack learning initiative and creativity, which leads to the problem of insufficient teaching quality. Therefore, this paper proposes a virtual exper...
详细信息
A new algorithm for constrained multi-objective optimization is presented. The algorithm treats the constraints as an objective and the immune clone and immune memory mechanism are introduced. Therefore, the new algor...
详细信息
A new algorithm for constrained multi-objective optimization is presented. The algorithm treats the constraints as an objective and the immune clone and immune memory mechanism are introduced. Therefore, the new algorithm could find the Pareto-optimal solutions from the feasible region and the edge of the infeasible region, which assures both the convergence and diversity of the obtained solutions. Simulation results show that the new algorithm has much better performance in finding a much better spread of solutions, in maintaining a better uniformity of the solutions and in obtaining a better convergence.
暂无评论