Estimation of Distribution Algorithm (EDA) is a new kinds of colony evolution algorithm, through counting excellent information of individuals of the present colony EDA construct probability distribution model, then sample the model produces newt generation. To solve the NP-Hard question as EDA searching optimum network structure a new Maximum Entropy Distribution Algorithm (MEEDA) is provided. The algorithm takes Jaynes principle as the basis, makes use of the maximum entropy of random variables to estimate the minimum bias probability distribution of random variables, and then regard it as the evolution model of the algorithm, which produces the optimal/near optimal solution. Then this paper presents a rough programming model for job shop scheduling under uncertain information problem. The method overcomes the defects of traditional methods which need pre-set authorized characteristics or amount described attributes, designs multi-objective optimization mechanism and expands the application space of a rough set in the issue of job shop scheduling under uncertain information environment. Due to the complexity of the proposed model, traditional algorithms have low capability in producing a feasible solution. We use MEEDA in order to enable a definition of a solution within a reasonable amount of time. We assume that machine flexibility in processing operations to decrease the complexity of the proposed model. Muth and Thompson’s benchmark problems tests are used to verify and validate the proposed rough programming model and its algorithm. The computational results obtained by MEEDA are compared with GA. The compared results prove the effectiveness of MEEDA in the job shop scheduling problem under uncertain information environment.

Grid computing is the technology used in building Internet-wide computing environment and integrating distributed and heterogeneous resources. However, in Grid environment, job scheduling is confronted with a great challenge. This paper focuses on lightweight jobs scheduling in Grid Computing. An Adaptive Fine-grained Job Scheduling (AFJS) algorithm is proposed. Compared with other fine-grained job scheduling algorithms based on grouping, AFJS can outperform them according to the experimental results. And some other related problems are also illustrated.

Recently, researches have shown that today'sbest-effort Internet with "one fits all" principle comes to an impasse. Addressed to the different QoS requirements oftoday's network services, we propose a new network architecture based on network virtualization technology for improving Internet QoS. It divides a “thick” network into multiple “thin” virtual networks deploying on the same substrate, each is customized with a special designing goal and runs a customized protocol. Then the traffic with different QoS requirements is classified at the ingress router,and distributed to different virtual networks, which are the
most suitable for carrying the special traffic. Also, we can deploy a service on multiple virtual networks, and making them working collaboratively to achieve a better QoS. To verify this idea, a prototype is implemented in LAN-scale network. By some simply designed experiments for comparison, we observe that, by use of our network architecture, service provider with specific QoS requirement can take its choice to choose appropriate virtual networks to achieve better QoS performance.

In order to improve the monitoring and controlling performances in construction projects, especially in multi-story projects, This paper researched the approaches to track the materials and components on construction project. A locating model in 3 dimensional spaces is proposed. The mentioned locating algorithms are compared, and shortages and advantages of each are given. The purpose is to find an optimal solution to manage the location information in construction project site with lower costs, higher locating accuracy and ideal reliability.

Hackers’ attacks are more and more intelligent, which makes it hard for single intrusion detection methods to attain favorable detection result. Therefore, many researches have carried out how to combine multiple security measures to provide the network system more effective protection. However, so far none of those methods can achieve the requirement of the practical application. A new computer information security protection system based on data fusion theory is proposed in this paper. Multiple detection measures are “fused” in this system, so that it has lower false negatives rate and false positive rate as well as better scalabilities and robust.

In wireless mesh networks, each node can be equipped with multiple network interface cards tuned to different channels. In this paper, we study the problem of collision-free multicast in multi-interface multi-channel wireless mesh networks. The concept of interface redundancy is proposed as a new criterion for the multicast/broadcast redundancy in wireless mesh networks, and we prove that building a multicast/broadcast tree with the minimum interface redundancy is NP-hard. We also prove that the minimum-latency multicasting problem in multi-channel wireless mesh networks is NP-hard. We present two heuristic-based algorithms which jointly reduce the interface redundancy and the multicast latency. Since broadcast can be considered as a special case of multicast, an approximate algorithm for low-redundancy broadcast tree construction is also proposed, which has a constant approximation ratio. Finally, the simulation results prove the effectiveness of our approach.

The quasi-regular patterns generated from hamiltonian has unique balanced symmetry. The patterns are quite suitable for the textile fabric. However the number of patterns just generated from hamiltonian are limited. Therefore in this paper a 110 of functions built by the methods of hamiltonian of superposition, nesting, combination and parameter variations, and it can increase the type and quantity. All of functions can generate fine patterns suited for textile fabric after computer graphics.
The system of textile pattern draft is generated by Visual basics for the computer. The system allows individual selection of function and parameters to change the pattern and the color conditioning. In this way, it successfully achieves the purpose.