Previous algorithms of data partitioning methods (DPMs) to find the exact K-nearest neighbors (KNN) at high dimensions are outperformed by a linear scan method [J.M. Kleinberg, Two algorithms for nearest neighbor search in high dimensions, 29th ACM Symposium on Theory of computing, 1997; R. Weber, H.-J. Schek, S. Blott. A quantitative analysis and performance study for similarity-search methods in high-dimensional spaces. in: Proc. of the 24th VLDB, USA, 1998]. In this paper, we present a “plug&search” method to greatly speed up the exact KNN search of existing DPMs. The idea is to linearize the data partitions produced by a DPM, rather than the points themselves, into a one-dimensional array-index, that is simple, compact and fast. Unlike most DPMs that support KNN search, which require storage space linear, or exponential [J.M. Kleinberg, Two algorithms for nearest neighbor search in high dimensions, 29th ACM Symposium on Theory of computing, 1997; M. Hagedoom, Nearest neighbors can be found efficiently if the dimension is small relative to the input size, ICDT 2003], in dimensions, the array-index requires a storage space that is linear in the number of mapped partitions. 相似文献
This paper presents a comparison between the live load distribution factors of simple span slab-on-girders concrete bridges based on the current AASHTO-LRFD and finite-element analysis. In this comparison, the range of applicability limits specified by the current AASHTO-LRFD is fully covered and investigated in terms of span length, slab thickness, girder spacing and longitudinal stiffness. All the AASHTO-PCI concrete girders (Types I–VI) are considered to cover the complete range of longitudinal stiffness specified in the AASHTO-LRFD. Several finite-elements linear elastic models were investigated to obtain the most accurate method to represent the bridge superstructure. The bridge deck was modeled as four-node quadrilateral shell elements, whereas the girders were modeled using two-node space frame elements. The live load used in the analysis is the vehicular load plus the standard lane load as specified by AASHTO-LRFD. The live load is positioned at the longitudinal location that produced the extreme effect, and then it is moved transversely across the bridge width in order to investigate all possibilities of one-lane, two-lane and three-lane design loads. A total of 886 bridge superstructure models were built and analyzed using the computer program SAP2000 to perform this comparison. The results of this study are presented in terms of figures to be practically useful to bridge engineers. This study showed that the AASHTO-LRFD may significantly overestimate the live load distribution factors compared to the finite-element analysis. 相似文献
A new method to achieve steady-state and dynamic-tracking desorption of organic compounds from activated carbon was developed and tested with a bench-scale system. Activated carbon fiber cloth (ACFC) was used to adsorb methyl ethyl ketone (MEK) from air streams. Direct electrothermal heating was then used to desorb the vapor to generate select vapor concentrations at 500 ppmv and 5000 ppmv in air. Dynamic-tracking desorption was also achieved with carefully controlled yet variable vapor concentrations between 250 ppmv and 5000 ppmv, while also allowing the flow rate of the carrier gas to change by 100%. These results were also compared to conditions when recovering MEK as a liquid, and using microwaves as the source of energy to regenerate the adsorbent to provide MEK as a vapor or a liquid. 相似文献
Distributed uplink scheduling and rate control in CDMA networks are considered in the case of adaptive antenna arrays present at the base station. The system model with omnidirectional antennas is generalized to the case where adaptive antenna arrays are deployed. Rate control in a probabilistic manner is investigated. Long-term control by the base station through token bucket constraints is incorporated in the system. Monte Carlo simulation results show considerable improvement when adaptive antenna arrays are used. Optimization of the rate transition probabilities is treated in the special case of on–off scheduling. 相似文献
Fully autotrophic nitrogen removal processes, such as the combined SHARON-Anammox process, help to improve the sustainability of wastewater treatment. Successful operation of such a completely autotrophic system is, among others, based on the strict control of the SHARON reactor in order to produce an Anammox-suited influent with a 1:1 ammonium:nitrite ratio. The high quality and high frequency measurements provided by a titrimetric set-up measuring the total ammonium (TAN) and total nitrite (TNO2) concentrations facilitate this control considerably. In this study, the use of a titrimetric set-up for monitoring the combined SHARON-Anammox process is investigated. The technique that interprets on-line collected titration curves was applied to a lab-scale system. Comparison with classic colorimetric results gave statistically indistinguishable results for TAN and TNO2 concentrations in the SHARON reactor. In the Anammox reactor, only TAN could be determined by the investigated method due to the very low TNO2 concentrations. Phosphate, a potential inhibitor of the Anammox process, is available as an additional measurement in the effluent of the SHARON reactor. Three measurements are thus combined in one single instrument. The proposed measuring technique holds different advantages over the other TAN and TNO2 measurement techniques such as on-site availability, easy automation, the absence of the need for high dilutions and cost reduction. 相似文献
The nature of disturbed thinking is explained from two different psychological perspectives and six major types of thought disorder are presented, classified, and discussed. The overlap between these different types of thought disturbance and some possible functions of disturbed thinking are outlined. A heuristic model for detecting thought disorder is presented; and, in addition, some complex conceptual and research issues involved with investigating and systematizing knowledge about disturbed thinking are explained. 相似文献
International Journal of Mechanics and Materials in Design - A generalized non-local stress–strain gradient theory is presented using fractional calculus. The proposed theory includes as a... 相似文献
The design and sustainability of reinforced concrete deep beam are still the main issues in the sector of structural engineering despite the existence of modern advancements in this area. Proper understanding of shear stress characteristics can assist in providing safer design and prevent failure in deep beams which consequently lead to saving lives and properties. In this investigation, a new intelligent model depending on the hybridization of support vector regression with bio-inspired optimization approach called genetic algorithm (SVR-GA) is employed to predict the shear strength of reinforced concrete (RC) deep beams based on dimensional, mechanical and material parameters properties. The adopted SVR-GA modelling approach is validated against three different well established artificial intelligent (AI) models, including classical SVR, artificial neural network (ANN) and gradient boosted decision trees (GBDTs). The comparison assessments provide a clear impression of the superior capability of the proposed SVR-GA model in the prediction of shear strength capability of simply supported deep beams. The simulated results gained by SVR-GA model are very close to the experimental ones. In quantitative results, the coefficient of determination (R2) during the testing phase (R2 = 0.95), whereas the other comparable models generated relatively lower values of R2 ranging from 0.884 to 0.941. All in all, the proposed SVR-GA model showed an applicable and robust computer aid technology for modelling RC deep beam shear strength that contributes to the base knowledge of material and structural engineering perspective.