Applied Mechanics and Materials Vols. 556-562

Abstract: In view of traffic queuing question, this article has established the dynamic queuing theory model, considered the waiting the traffic flow magnitude and the vehicles density. On this basis, through computer random simulations,we found that traffic conditions can be divided into three statuses,and found out the key conditions of formation for each status. The results can be widely promoted, and can provide advice to government

Abstract: Speech signal can be effect by the external interruption in the process of flight simulation, which results in the deteriorated speech processing performance in flight simulation system. To solve the problem, speech enhancement module has been developed using the spectral subtraction algorithm based on the flight simulation system developed by the research team. The module will improve the recognition rates and provide advanced anti-jam capabilities for the speech processing performance of the flight simulation system. The feasibility and validity of this speech enhancement module is validated by processing the interrupted speech signal. The work in this paper has laid foundation for the application of the speech enhancement technology in flight simulation system.

Abstract: In the iris recognition, as the texture feature of iris is complicated , and it is affected by the acquisition environment, time, condition and other factors, so it is difficult to express its complete information in single features of iris. Therefore, the combination of a variety of features or integration of expression and recognition method becomes a new kind of choice. Information integration concept originated in the 1970s. In recent years, because of the gradually in-depth recognition of people, the application of information integration technology in the pattern recognition becomes a hot spot of the research, especially when the single pattern cannot complete various aspects of requirements well in the pattern recognition, the adoption of integration technology can integrate various aspects of information provided by multi-pattern features, so as to obtain more comprehensive and accurate information of the recognized objects and overcome the limitations of recognition method through single-pattern feature, thus getting a satisfactory effect.

Abstract: the paper suggests a kind of self-adapted layer contrast enhancement algorithm for medical images, which, in reference to Laplacian pyramid function, could compose images, and enhance pixel contrast on each layer through F/E processing and use of self-adaptation Sigmoid function, and it uses layer enhancement factor to control the enhancement degree. The result shows that details of medical images can be displayed clearly with this algorithm without selecting characteristic center dimensions. It will not generate artifacts and thus improves visual effect of medical images.

Abstract: Path planning is the core issues in the artificial intelligence field of games, and how to establish an effective method of path planning is still focused on. A new algorithm based on both the benefits of global path planner methods and local path planner methods and those of A star algorithm and improved artificial potential field method is proposed for NPC path planning in web games. Its feasibleness and effectiveness are also demonstrated by simulation results.

Abstract: Along with Web on information content sharp increasing, Web data mining function obviously important. This article analyzed the Web-based data mining’s definition and its classification and its process as well as its application.

Abstract: Network load is balanced and energy is in stable and sustainable is the important guarantee for the stability of the network. Traditional clustering algorithm for ad hoc network can effectively achieve a single point of network stability, and there is no efficient algorithm for long-term stability. Improvements have been made in this context traditional network clustering algorithm, based on efficient cluster head node using residual energy and communications networking sensor networks nodes remaining weaknesses in the energy networks and strengthening, improving the overall balance and effectiveness of the network, and ultimately to improve the long-term stability in wireless ad hoc networks sensor networks.

Abstract: Effective task scheduling is crucial for achieving high performance in heterogeneous computing environments. Whiling scheduling Out-Tree task graphs, many previous heterogeneity based heuristic algorithms usually require high scheduling costs and may not deliver good quality schedules with lower costs. Aiming at the characteristics of Out-Tree task graphs and the features of heterogeneous environments and adopting the strategy based on expected costs and task duplications, this paper proposes a greedy scheduling algorithm, which, at each scheduling step, tries to guarantee not to increase the schedule length, schedules the current task onto the used processor which minimizes its execution finish time; meanwhile, takes load balances into account to economize the use of processors. The comparative experimental results show that the proposed algorithm has higher scheduling efficiency and robust performance, which could produce better schedule which has shorter schedule length and less number of used processors.

Abstract: Online advertising budget analysis and optimization of delivery systems, provide a modern service model with innovation and intelligence for Chinese advertising market. This paper describes functions and use case diagrams of this system,as well as its realizing method.

Abstract: By using rice output data of Baise statistical yearbook from 1989 to 2010, based on the Grey theory and Grey forecast models, which are GM(1,1), have been adopted to predict the rice output of Baise city in this paper. So we get the GM(1,1) prediction model and predict the rice production of Baise city from 2011 to 2020. The results show that the Grey theory model in predicting rice output is feasibility and reasonable.