The subject of this work is optimization of outputs of L-band high power microwave tubes. These outputs are constructed as coaxial-to-waveguide transitions with a vacuum barrier in a form of glass or ...ceramic cup. The goal of optimization is to obtain sufficiently low reflection loss in the predefined frequency band and to avoid so called hot spots caused by excessive dissipation of microwave power in parts of the cup. Electromagnetic simulator has been applied to model the behavior of the optimized transition and to propose its optimum shape. The proposed solutions were verified by Z.E. Lamina SA in prototypes of high power (pulsed 600 kW) amplitrons and are to be used in manufacturing practice.
•Cooperative game is introduced into network DEA models to depict the internal complex relationships.•In this three-stage framework, shared inputs and undesirable outputs are ...considered.•Super-efficiency is applied for further differentiation and ranking of the multiple efficient DMUs.•Tobit model is used to investigate the factors associated with low efficiency.
In this paper, the R&D green innovation of the Chinese high-tech industry can be regarded as a system of three consequent stages. Firstly, considering the existence of multiple effective decision-making units at the same time, a three-stage super-efficiency DEA model is proposed. Secondly, considering the “internal relationships” and “the distribution structure of initial inputs and undesirable outputs”, the above model is extended into a three-stage super-efficiency DEA model with a cooperative game based on shared inputs and undesirable outputs. Thirdly, to effectively obtain the unique optimal efficiency of each stage, an efficiency decomposition approach is proposed. Finally, an empirical example is conducted to evaluate the overall efficiency and the individual efficiency of R&D green innovation. Sensitivity analysis is conducted to analyze the reasonable allocation information of initial inputs. Furthermore, the Tobit model is adopted to analyze the main influencing factors of R&D green innovation. The empirical results show that: (1) Overall efficiency of the R&D green innovation is low and there exists much space for improvement. (2) Provincial development is uneven, and there are obvious differences in efficiency among the three stages within each province. (3) Technology market development, the degree of dependence on foreign trade, the degree of industrial structure, foreign direct investments have significant effects on the R&D green innovation; while government funding and R&D investment intensity have not produced obvious effects on the R&D green innovation.
Probabilistic forecasting of solar photovoltaic (PV) generation is critical for stochastic or robust optimisation-based power system dispatch. This study proposes a randomised learning-based hybrid ...ensemble (RLHE) model to construct the prediction intervals of probabilistic PV forecasting. Three different randomised learning algorithms, namely extreme learning machine, randomised vector functional link, and stochastic configuration network, are ensembled as a hybrid forecasting model. Besides, bootstrap is used as the ensemble learning framework to increase the diversity of training samples. For each algorithm, a decision-making rule is designed to evaluate the credibility of the individual outputs and the incredible ones are discarded at the output aggregation step. The weight coefficients of the aggregated outputs of the three algorithms are then optimised to compute the final point forecast results. Based on the point forecast results, the prediction intervals are constructed considering both model misspecification uncertainty and data noise uncertainty. The variance in model misspecification uncertainty is directly calculated with the individual outputs and the variance in data noise uncertainty is separately trained with an RLHE model. The proposed method is tested with an open dataset and compared with several benchmarking approaches.
•A three-stage framework is constructed to identify the internal stages of the banking system.•Carryovers are considered at the capital allocation stage to connect the neighboring periods.•Triangular ...type-2 fuzzy numbers are employed to characterize NPLs to determine the sensitivity of the banks to credit risk.•Shared inputs are considered in the modeling to enable managers to optimally determine the appropriate proportions for the reallocation of resources in the different stages.•Internal resource imbalances are considered to allow some flexibility in the common variables between the different stages and periods.
Efficiency evaluations are vital for banks so that they can determine their future development and enhance their competitiveness. To comprehensively explore a bank’s internal structures and identify the specific reasons for any inefficiencies, three stages of banking systems need to be examined; capital organization, capital allocation, and profitability. To measure the efficiencies over consecutive periods, this paper developed a multi-period, multi-stage DEA model, in which unused assets were carried over to subsequent periods, fixed assets and employee salaries were regarded as shared inputs for all three stages, and non-performing loans, which were characterized using triangular type-2 fuzzy numbers, were introduced as undesirable outputs to reflect credit risk. The developed model was applied to a case study to evaluate the efficiencies of listed Chinese commercial banks from 2014 to 2016, from which a disparity in efficiencies was found; that is all banks were found to be generally inefficient; however, the inefficiencies occurred in different stages for different types of banks. Varying optimistic-pessimistic attitudes were applied to identify the overly sensitive banks and comparisons were conducted to provide managerial insights and verify the superiority of the proposed model. It was concluded that to enhance overall efficiency, banks need to have a reasonable business scale, that adopting a three-stage analytical framework can better identify efficiencies and the weaker stages, and that neglecting carryovers can overestimate bank performance.
Object skeletons are useful for object representation and object detection. They are complementary to the object contour, and provide extra information, such as how object scale (thickness) varies ...among object parts. But object skeleton extraction from natural images is very challenging, because it requires the extractor to be able to capture both local and non-local image context in order to determine the scale of each skeleton pixel. In this paper, we present a novel fully convolutional network with multiple scale-associated side outputs to address this problem. By observing the relationship between the receptive field sizes of the different layers in the network and the skeleton scales they can capture, we introduce two scale-associated side outputs to each stage of the network. The network is trained by multi-task learning, where one task is skeleton localization to classify whether a pixel is a skeleton pixel or not, and the other is skeleton scale prediction to regress the scale of each skeleton pixel. Supervision is imposed at different stages by guiding the scale-associated side outputs toward the ground-truth skeletons at the appropriate scales. The responses of the multiple scale-associated side outputs are then fused in a scale-specific way to detect skeleton pixels using multiple scales effectively. Our method achieves promising results on two skeleton extraction datasets, and significantly outperforms other competitors. In addition, the usefulness of the obtained skeletons and scales (thickness) are verified on two object detection applications: foreground object segmentation and object proposal detection.
lCultivated land has the dual effect of suppling agricultural products and releasing a large amount of carbon emissions.lAn evaluation framework considering carbon emissions resulting from cultivated ...land use has been established.lSpatial disparities of cultivated land use efficiency and its influencing factors were further analyzed.lEffective ways should be taken to alleviate the disparities of cultivated land use efficiency in China.
Rapid urbanization and industrialization has worsened the situation of the scarce cultivated land resources of China. It's therefore of great importance for sustainable development based on the systematic evaluation on cultivated land use efficiency (CLUE). This study took carbon emissions resulting from cultivated land use into the measurement framework of CLUE, and a slack-based measure (SBM) model with undesirable outputs, boxplot, kernel density estimation and Tobit regression model are adopted for the analysis of 31 provinces in China from 2000 to 2017. The results showed that there was an increasing trend in CLUE in China from 0.5236 in 2000 to 0.8501 in 2017, with the growth rate of 38.40%. Most of provinces in China have much lower levels of CLUE with significantly spatial disparities. In particular, Hainan, Chongqing, Sichuan and Guizhou are always most efficient with the highest value of 1. At the regional level, the average value of CLUE in the northeastern region is the highest, followed by the western, eastern and central regions, and the CLUE in the eastern region is more unstable than the other three regions. The results of Tobit regression show that natural conditions, cultivated land resource endowments, agricultural production conditions, regional economic development and regional science and technology development are important factors resulting in the disparity of China's CLUE.
Economic growth in China needs lots of energy to support, but how to obtain a sufficient energy supply is now a problem. Meanwhile, environmental pollution caused by energy consumption is a problem ...that urgently needs to be solved. In order to adapt China’s fast developing economy, and on the basis of current policies, paying more attention to energy saving and environmental protection is more important. Therefore, to reasonably analyze China’s regional environmental efficiency and its determinants has practical significance. This article uses the Super-SBM model under undesirable outputs to measure regional environmental efficiency in China, and then explores influential factors of China’s environmental efficiency between 1991–2001 by means of the Tobit regression model. The empirical results show that during the sample period, the overall average level of China’s environmental efficiency is low, and the gap between different provinces and areas is large. The eastern area’s efficiency is obviously superior to the central and western area, which shows the necessity at present to optimize regional integration and promote innovation of energy conservation and emission reduction technology. At the same time, the “Eleven-Five ” year plan clearly states the energy saving and consumption reducing index, which motivates a great increase in China’s overall environmental efficiency, and promotes the fast development of resource conservation and environmentally friendly societal construction. The analysis of the influence factors by the Tobit model shows that fiscal decentralization and technology progress can increase environmental efficiency on the whole while economic scale and regional difference can also influence the efficiency. However, these factors’ influence is not the same in different areas. Finally, based on the research results, this paper gives policy recommendations on reducing pollution emissions and increasing China’s environmental efficiency.
•We conduct an extended survey of DEA-based approaches that include bad outputs.•The widely-used weak disposability assumption presents important shortcomings.•Approaches considering multiple ...frontier technologies are more promising.
This article is a critical review of methods integrating environmental aspects into productive efficiency. We describe the classic modelling approach relying on the weak disposability assumption, and explain the major recent developments around the inclusion of undesirable outputs in production technology modelling, namely the materials balance principles and the weak G-disposability, the by-production modelling and the cost disposability assumption, and the unified model under natural and managerial disposability concepts. We discuss the limits inherent in each methodology and suggest future research perspectives.
Gradient-boosted decision trees (GBDTs) are widely used in machine learning, and the output of current GBDT implementations is a single variable. When there are multiple outputs, GBDT constructs ...multiple trees corresponding to the output variables. The correlations between variables are ignored by such a strategy causing redundancy of the learned tree structures. In this article, we propose a general method to learn GBDT for multiple outputs, called GBDT-MO. Each leaf of GBDT-MO constructs predictions of all variables or a subset of automatically selected variables. This is achieved by considering the summation of objective gains over all output variables. Moreover, we extend histogram approximation into the multiple-output case to speed up training. Various experiments on synthetic and real data sets verify that GBDT-MO achieves outstanding performance in terms of accuracy, training speed, and inference speed.