We consider the problem of maximizing the lifetime of a given multicast connection in a wireless network of energy-constrained (e.g., battery-operated) nodes, by choosing ideal transmission power levels for the nodes ...
详细信息
We consider the problem of maximizing the lifetime of a given multicast connection in a wireless network of energy-constrained (e.g., battery-operated) nodes, by choosing ideal transmission power levels for the nodes relaying the connection. We distinguish between two basic operating modes: In a static power assignment, the power levels of the nodes are set at the beginning and remain unchanged until the nodes are depleted of energy. In a dynamic power schedule, the powers can be adjusted during operation. We show that while lifetime-maximizing static power assignments can be found in polynomial time, for dynamic schedules the problem becomes NP-hard. We introduce two approximation heuristics for the dynamic case, and experimentally verify that the lifetime of a dynamically adjusted multicast connection can be made several times longer than what can be achieved by the best possible static assignment.
We address abstraction in the setting of probabilistic reactive systems, and study its formal underpinnings for the strictly alternating model. In particular, we define the notion of branching bisimilarity and study i...
详细信息
This paper proposes a kind of theory architecture model of mobile Internet service management and proposes a high-level efficiency service discovery algorithm based on this theory architecture model. An emulating test...
详细信息
This paper proposes a kind of theory architecture model of mobile Internet service management and proposes a high-level efficiency service discovery algorithm based on this theory architecture model. An emulating test is designed. The result shows: by two-phase process, it not only saves much wireless bandwidth resource, but also enhances service availability and enhances service discovery efficiency.
In the encoding algorithm of H.264, each possible coding mode should be checked to choose the best one for each part of the picture, thus make the process of mode-decision rather comp.ex. In this paper, a fast mode de...
详细信息
In the encoding algorithm of H.264, each possible coding mode should be checked to choose the best one for each part of the picture, thus make the process of mode-decision rather comp.ex. In this paper, a fast mode decision algorithm is proposed to reduce the comp.exity of mode decision procedure in H.264 encoder., in which possible modes are judged beforehand based on the temporal and spacial correlation and the whole procedure of mode decision is simplified then. The results show that the algorithm can increase the encoding speed by more than 20% at most with little degradation in rate-distortion performance.
In order to solve the problem that the traditional point-matching algorithms have relative high time comp.exity and are easy to converge into local optima, a registration method for feature point sets based on Voronoi...
详细信息
In order to solve the problem that the traditional point-matching algorithms have relative high time comp.exity and are easy to converge into local optima, a registration method for feature point sets based on Voronoi diagram and evolution strategy is presented. In the new method, the mean square of Euclidean distances of corresponding points is regarded as the optimal object. By using the new given area-filling algorithm to generate the Voronoi diagram of the reference image, which divides the reference image area into several non-overlapping sub-areas and each of them contains and only contains one feature point of the reference image. For any point in a given sub-area, the closest feature point of the reference image is just the one in the same sub-area. According to the Voronoi diagram's information, the objective function can be iteratively calculated at a high speed by combination of SVD decomp.sition and evolution strategy in the process of iteration. In the experiment of multimodality medical image registration, comp.red with the traditional ICP algorithm, the result shows that the proposed method has higher speed and better ability to avoid local optima.
A high order accurate difference scheme is presented for nonlinear hyperbolic conservation laws. Firstly, the comp.tational domain is divided into many pieces of non-overlapping subdomains, and then each subdomain is ...
详细信息
A high order accurate difference scheme is presented for nonlinear hyperbolic conservation laws. Firstly, the comp.tational domain is divided into many pieces of non-overlapping subdomains, and then each subdomain is further subdivided into equal small-cells according to required accuracy. Secondly, by the flow direction, flux splitting will be used, and the positive/negative numerical fluxes are reconstructed at small-cell boundaries, and the correction are introduced. With the aid of high-order total variation diminishing (TVD) Runge-Kutta time discretization, a high-order, high resolution, conservative difference scheme is constructed for one dimensional nonlinear hyperbolic conservation laws. The scheme is extended to two dimensional hyperbolic conservation equation by dimension-by-dimension method, and the maximum and minimum bounded (MmB) property of the scheme is proved. Furthermore, The extension to two dimensional system of conservation laws is carried out. Finally, several numerical experiments on two dimensional Euler equations are given, and numerical results are satisfied.
A high-order, high resolution, conservative TVD difference scheme is presented for one dimensional hyperbolic conservation equations. The basic idea is as follows. Firstly, the comp.tation domain is divided into many ...
详细信息
A high-order, high resolution, conservative TVD difference scheme is presented for one dimensional hyperbolic conservation equations. The basic idea is as follows. Firstly, the comp.tation domain is divided into many non-overlapping subdomains, and then each subdomain is further subdivided into small cells according to the required accuracy;Secondly, by the flow direction, flux splitting is introduced, and high-order approximation in the subdomain are used to comp.te the positive/negative numerical fluxes at cell boundaries. Furthermore, TVD corrections are considered to prevent osci.lations near discontinuities from the high-order interpolation. Moreover, by means of high-order TVD Runge-Kutta time discretization, a high-order fully discretization method is obtained. The extension to one dimensional systems is also carried out. Finally, numerical experiments on one dimensional Euler equations are given, and numerical results are satisfactory.
Kaisa Nyberg and Takashi Satoh gave some methods for constructing multiple output bent functions in paper[1,2] respectively, while these bent functions belong to Maiorana-McFarland bent functions. In this paper, an op...
详细信息
Kaisa Nyberg and Takashi Satoh gave some methods for constructing multiple output bent functions in paper[1,2] respectively, while these bent functions belong to Maiorana-McFarland bent functions. In this paper, an optimizing method for multiple output bent functions in paper[1,2] is presented by using Cryptographic functions' input-variety comp.unding reversible transform. As a result, many new multiple output bent functions were obtained, and these functions have following properties: (1) they do not belong to Maiorana-McFarland bent functions;(2) they have highest algebraic degree and maximal number of output bits;(3) they also have good construction number. Moreover, it also means that comp.unding reversible transform is very useful for constructing good Cryptographic functions and the method is concise and easily implemented.
Firstly, the image was decomp.sed by wavelet transform to multiresolution image. Multiresolution image pyramid was constructed by multiresolution image. Secondly, the noise of image was better reduced in wavelet trans...
详细信息
Firstly, the image was decomp.sed by wavelet transform to multiresolution image. Multiresolution image pyramid was constructed by multiresolution image. Secondly, the noise of image was better reduced in wavelet transform domain. Finally, the image was segment by classification method of Kohonen neural network that increases speed and accuracy of classification.
暂无评论