Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2019-05-10 Cooperative journals: 《计算机应用研究》
Abstract: In order to avoid local-trap and premature convergence of robot motion planning in 2D complex environments, this paper proposes an improved metaheuristics-adaptive genetic algorithm (GA) . Firstly, it adoptes the random Dijkstra algorithm to create initial population; secondly, in each generation of the GA, improves the created paths, and replaces the conventional selection operator in GA with an adaptive one; finally, by using feedback information of the search process, the adaptive selection operator can control the selective pressure appropriately throughout the algorithm. To validate the effectiveness of the proposed method, compares the algorithm with two other methods in MATLAB. The results show that the proposed method can avoid the local convergence problem in motion planning, and can generate feasible path in complex environments.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2019-04-01 Cooperative journals: 《计算机应用研究》
Abstract: Aiming at the problem that the existing attitude estimation method can not accurately extract the feature parameters, this paper proposed a hierarchical model based on structured support vector machine (SSVM) and convolutional neural network (CNN) . First, it showed how a SSVM based on the PS component model could be implemented as a two-layer neural network, where the first layer was the convolutional layer and the other layer was the loss-enhanced inference layer. Then, by transforming the structured form of the model into a neural network in the model, the proposed method could simultaneously learn the structural model and the appearance model, and then backpropagated the error to learn the underlying learnable parameters. These parameters could be derived from the appearance model features. Extracted out. Finally, the SSVM model was transformed into a neural network model, the error was propagated back to the lower layer, and the exact SSVM loss was calculated, while the original SSVM was learned by the sub-gradient-based method. Comparing the model with the current advanced recognition model, the results show that the proposed success rate of the hierarchical model is 6% higher than the comparison method and has stronger recognition performance.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2019-01-03 Cooperative journals: 《计算机应用研究》
Abstract: An efficient LFA implementation method which is based on Incremental Shortest Path First Algorithm (ERPISPF) was proposed to reduce the computational overhead and deployment difficulty of the existing LFA algorithm. The paper first turns the problem of quick implementation of LFA into how to efficiently calculate the minimum cost of all its neighbors to all other nodes of the network on the shortest path tree rooted at the compute node. Then a theorem for calculating the cost is presented and its correctness is proved. Finally, the time complexity of the algorithm was theoretically analyzed. Experiments show that compared with LFA algorithm, ERPISPF not only has less computation overhead, but also provides the same failure protection rate as LFA.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-11-29 Cooperative journals: 《计算机应用研究》
Abstract: Reducing the energy consumption of the Internet has become an urgent problem. However, the existing energy efficient routing algorithms are facing the following two problems: a) all of them will degrade the network performance in varying forms, such as network congestion, routing flap, routing availability and uneven traffic distribution; b) all of them require real-time traffic information, which lead to high complexity of the algorithm. Therefore, this paper designs a green energy efficient scheme EEIPFRR which is based on IP fast reroute, taking into account the energy saving, network performance and algorithm complexity. The experimental results show that compared with DLF algorithm, it can not only reduce network energy consumption, but also have less path stretch, lower algorithm complexity and maximum link utilization.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-11-29 Cooperative journals: 《计算机应用研究》
Abstract: The existing routing protection schemes do not consider the importance of nodes in the network; however, the importance of different nodes in the network is not the same in real networks. To solve this problem, this paper proposed an intra-domain routing protection algorithm based on node diversity (RPBND) . Firstly, the computing node constructed a shortest path tree rooted at the destination node, which ensured the compatibility between the RPBND and the current deployment intra-domain routing algorithm. Then, it builded a directed acyclic graph on the basis of the shortest path tree, which maximized the Internet routing availability. The experiment results show that RPBND greatly improves the Internet routing availability and reduces the network disruption time caused by the faults, which provides the sufficient basis for ISPs to deploy the intra-domain routing protection scheme.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-07-23 Cooperative journals: 《计算机应用研究》
Abstract: Recently, open source software is popular in software industry. However, the reliability for open source software is widely questioned by people. How to assess the reliability of open source software is an important issue. Compared with traditional closed source software, two factors, i. e. fault introduction and the delay time between the fault detection and correction are necessarily considered when building software reliability model of open source software. In this paper, we propose software reliability models for open source software considering the correction process and imperfect debugging. We use the experiment to validate the goodness-of-fit and predictive performance of the proposed model using two fault data sets of open source software. The experimental results show that the proposed models have a good fitting and predictive power in the reliability evaluation of open source software. The proposed models can be used to evaluate the reliability of open source software in the real-world development of open source software.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-05-18 Cooperative journals: 《计算机应用研究》
Abstract: Software defined network (SDN) is a novel network architecture separating control plane and forwarding plane. SDN has been favored by the industry because of its flexibility and controllability. However, SDN usually employs the best paths to forward packets, which is difficult to deal with the node or link failures in the network. In order to improve the availability of SDN network, this paper proposed an intra-domain routing protection scheme based on SDN (RPBSDN) . The scheme could calculate multiple backup next hops for each source-destination, and guaranteed loop-freeness of the induced forwarding path by the underlying partial order of the nodes. The experimental results show that the scheme not only has small computational complexity, but also greatly improve the availability of the network.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-05-18 Cooperative journals: 《计算机应用研究》
Abstract: To solve the problem of parameters selection in the process of constructing the distributional representations of Chinese words, this paper performed a systematic study. Six kinds of parameters were selected for comparison experiments, and the quality of the distributional representations of Chinese words obtained under different parameter settings was evaluated on the Chinese semantic similarity task. The experimental results show that, by choosing appropriate parameters, the distributional representations of Chinese words can also get higher performance on the similarity task, Moreover, the quality of such high-dimensional distributional representations is even superior to low-dimensional word representations based on neural network or matrix factorization.
Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-04-19 Cooperative journals: 《计算机应用研究》
Abstract: Aiming at the drawbacks of the classical WSN coverage model, especially if a sensor dies, K-coverage model requires at least k sensor nodes to monitor whether there is a target within its coverage area. This paper proposed a mobile WSN coverage method based on improved adaptive genetic algorithm, which provided continuous monitoring of specified targets for longest possible time with limited energy resources. The algorithm took into account that the motion sensor could move at variable speeds continuously to ensure that all targets were within their coverage. Simulation results show that in the case of mobile nodes, the life span and the number of data packets of the improved method are obviously improved compared with other commonly models.