To solve the problems in face editing， such as unnatural editing results and great changes in generated images， a controllable face editing algorithm with closed-form solution was proposed. Firstly， n latent vectors were sampled randomly to construct a sample matrix， and the top k principal component vectors of the matrix were calculated. Then， five attributes of face image were obtained by ResNet-50， and the semantic boundary of each attribute was calculated by Support Vector Machine （SVM）. Finally， the interpretable direction vectors of these attributes were calculated， which were as closed to the principal components vectors as possible and stayed as far away from the semantic boundary of the corresponding attribute as possible at the same time， thereby reducing the coupling between facial attributes， and improving the controllability in face editing. Because the algorithm has a closed-form solution， it has high efficiency. Experimental results show that the compared with closed-form Factorization of latent Semantics in GANs （SeFa） algorithm and Discovering Interpretable Generative Adversarial Network Controls （GANSpace） algorithm， the proposed algorithm increases the Inception Score （IS） by 19% and 26% respectively， decreases the Fréchet Inception Distance （FID） by 4% and 37% respectively， and decreases the Maximum Mean Discrepancy （MMD） by 15% and 48% respectively. It can be seen that this algorithm has good controllability and decoupling.

Aiming at the problems that the deep spectral clustering models perform poorly in training stability and generalization capability， a Deep Spectral Clustering algorithm with L1 Regularization （DSCLR） was proposed. Firstly， L1 regularization was introduced into the objective function of deep spectral clustering to sparsify the eigen vectors of the Laplacian matrix generated by the deep neural network model. And the generalization capability of the model was enhanced. Secondly， the network structure of the spectral clustering algorithm based on deep neural network was improved by using the Parametric Rectified Linear Unit activation function （PReLU） to solve the problems of model training instability and underfitting. Experimental results on MNIST dataset show that the proposed algorithm improves Clustering Accuracy （CA）， Normalized Mutual Information （NMI） index， and Adjusted Rand Index （ARI） by 11.85， 7.75， and 17.19 percentage points compared to the deep spectral clustering algorithm， respectively. Furthermore， the proposed algorithm also significantly improves the three evaluation metrics， CA， NMI and ARI， compared to algorithms such as Deep Embedded Clustering （DEC） and Deep Spectral Clustering using Dual Autoencoder Network （DSCDAN）.

Group activity recognition is a challenging task in complex scenes， which involves the interaction and the relative spatial position relationship of a group of people in the scene. The current group activity recognition methods either lack the fine design or do not take full advantage of interactive features among individuals. Therefore， a network framework based on partitioned attention mechanism and interactive position relationship was proposed， which further considered individual limbs semantic features and explored the relationship between interaction feature similarity and behavior consistency among individuals. Firstly， the original video sequences and optical flow image sequences were used as the input of the network， and a partitioned attention feature module was introduced to refine the limb motion features of individuals. Secondly， the spatial position and interactive distance were taken as individual interaction features. Finally， the individual motion features and spatial position relation features were fused as the features of the group scene undirected graph nodes， and Graph Convolutional Network （GCN） was adopted to further capture the activity interaction in the global scene， thereby recognizing the group activity. Experimental results show that this framework achieves 92.8% and 97.7% recognition accuracy on two group activity recognition datasets （CAD （Collective Activity Dataset） and CAE （Collective Activity Extended Dataset））. Compared with Actor Relationship Graph （ARG） and Confidence Energy Recurrent Network （CERN） on CAD dataset， this framework has the recognition accuracy improved by 1.8 percentage points and 5.6 percentage points respectively. At the same time， the results of ablation experiment show that the proposed algorithm achieves better recognition performance.

A novel kernel-based non-negative sparse representation (KNSR) method was presented for face recognition. The contributions were mainly three aspects: First, the non-negative constraints on representation coefficients were introduced into the Sparse Representation (SR) and the kernel function was exploited to depict non-linear relationships among different samples, based on which the corresponding objective function was proposed. Second, a multiplicative gradient descent method was proposed to solve the proposed objective function, which could achieve the global optimum value in theory. Finally, local binary feature and the Hamming kernel were used to model the non-linear relationships among face samples and therefore achieved robust face recognition. The experimental results on some challenging face databases demonstrate that the proposed algorithm has higher recognition rates in comparison with algorithms of Nearest Neighbor (NN), Support Vector Machine (SVM), Nearest Subspace (NS), SR and Collaborative Representation (CR), and achieves about 99% recognition rates on both YaleB and AR databases.

For the issue of how to quickly estimate the accurate, reliable due date according to the order information and the features of the production system in Assembly To Order (ATO), a due date prediction model was constructed based on the influential mechanism analysis of the uncertainty factors. The model parameters included three parts: order release time, assembly cycle time and abnormal tardiness. Order release time was based on the availability of materials and production capacity. The assembly cycle time and abnormal tardiness were predicted by using Support Vector Regression (SVR) method based on actual production history data. The case study shows that the predicted results of the model are close to actual due date and it can be used to guide the order's delivery time consultation.

Due to the complexity and the dynamic changes of the coal mine environment, the concentrations of harmful gases are difficult to be accurately monitored. The traditional monitoring methods use a single sensor to pick-up information, and the collected data have simple data form, low reliability, big error and so on. Concerning these problems, a new method was proposed in this paper, that is, sampling a variety of heterogeneous gases sources, and then taking advantage of the strong classification algorithm to filter, lastly fusing the above obtained information. As experiments state, the new method significantly improve the reliability of the mine monitoring system.

In view of the kinematic model of mobile robot, a tracking controller of global asymptotic stability was proposed. The design of tracking controller was divided into two parts: The first part designed the control law of angular velocity by using global fast terminal sliding mode in order to asymptotically stabilize the tracking error of the heading angle; the second part designed the control law of linear velocity by using the Lyapunov method in order to asymptotically stabilize the tracking error of the planar coordinate. By combining Lyapunov stability theorem and two control laws, the mobile robot can track the desired trajectory in a global asymptotic sense when the angular velocity and the linear velocity satisfy these control laws. The experimental results show that the mobile robot can track desired trajectory effectively. It is helpful for promoting the practical application.

An improved artificial immune intrusion detection model is proposed based on ARTIS(Artificial Immune System)--a distributed intrusion detection model proposed by Hofmeyr. It aims to overcome defects of the existing artificial immune IDS. To improve the quality and reduce the scale of memory and mature detector, the improved model uses the protocol analysis technology to make co-stimulation of the immune module. The protocols are taken into account while generating and organizing detectors, so the inefficiency of traditional AIS can be covered. Weight based r-continuous matching rules are taken to improve matching accuracy of the antibody-antigen reactions. Meanwhile, the co-stimulation module can automatically generate dynamic filter rules for firewall when Flood attack occurs. Finally, we have a simulation test and comparative analysis on improved model and ARTIS model by using DARPA data sets owned by MIT Lincoln lab and the results evaluate the feasibility and effectiveness of our improved model.

Since it is difficult to determine the network boundaries and topology is very flexible in Wireless Mesh Network (WMN), topology information collection and reconstruction have great delay, so that real-time WMN monitoring accuracy can not be ensured. This paper proposed a real-time monitoring method based on improved A algorithm for the topology state of WMN to get the real-time state and give out response to abnormity. Through limiting the path length, reducing the search scope and adding the number of repeated searched edges to heuristic of A, the method solved the problem that path may be recovered and too long for topology real-time monitoring. The simulation results show that compared with the original algorithm, the improved algorithm has a higher speed in convergence, and it can update the topology construction in shorter time.

It is necessary to store massive real-time data into database and query records from database in real-time on the field of satellite ground device monitoring. Taking account of the characteristics of real-time data and Judy array, a bitmap memory allocation method based on memory map file was proposed. A hybrid index mechanism which employed Hash table, B+ tree and Judy array was designed. Through insertion and querying of massive records, the experimental results show that bitmap allocation method avoids the generation of massive tiny memory holes. Being combined with bitmap allocation method, the hybrid index mechanism provides real-time index insertion and record querying for applications.

To solve the problem of transmitting credentials and other resources through unsafe physical channels during an Automated Trust Negotiation (ATN), a transmission scheme for credentials and resources was proposed based on magic cube algorithm. Through the magic cube algorithm, a transformation sequence was formed in terms of the request or the resource of negotiation initiator, followed by the digital digest to generate the information transformation sequence. According to the logical expression composed of credentials which represent the condition negotiation success, the information transformation sequence was shuffled to form an information transmission sequence, which was sent to the negotiation receiver. The information transmission sequence was reciprocally transformed by the negotiation receiver according to his own credentials. This scheme has many features of the one-round credential exchange, and little network cost. The example shows that the scheme is feasible, and the experimental results show that the scheme has good security and efficiency and low information transmission capacity.

Because of the user's mobility and the location dependency of data, new challenge has been brought to cache replacement strategy for Location Dependent Query (LDQ). Based on the detailed analysis of the space location characteristics of Location Dependent Data (LDD) and several typical replacement strategies of location dependent cache, the authors proposed a prioritized approach cache replacement based on the lowest access cost (PLAC), the PLAC took some important factors into account such as access probabilities, update rates, data distance, valid scope. To ensure the maximum utilization of limited cache, the PLAC cache replacement strategy decided which data would be replaced according to the value of the lowest cost function. The contrast experiments show that the PLAC increases cache hit rate and shortens query average response time more effectively than other location dependent cache replacement strategies.

On the basis of analyzing the mixed Client/Server (C/S) and Browser/Server (B/S) architecture pattern and AJAX technology, a novel mixed architecture pattern was proposed, which can unify the foreground interaction method of B/S and C/S and make the servers share effectively, thus enhance the scalability and maintainability of the system. According to the proposed pattern, the architecture of the spacecraft dynamical application platform was given, as a reference to the system with similar architecture.

Considering the demand of the date relationship and service request multiform based on XML documents, this paper proposed a new frequent pattern tree algorithm for selected incremental vector items set of refactoring XML structure (XFP-tree). Bases on the XML Key, the algorithm firstly dealt with XML structure to vector matrix, then used project frequent pattern tree to optimize the XML structure through dissociating, uniting, updating and canceling to satisfy the conciseness of the XML structure and query multiversity. Combining project and tree-structure manipulation, this paper discussed the dividing rule of xml key vector matrix frequent pattern. This rule improved the algorithm efficiency by establishing heuristic strategy and support thresholds. Contrasted with other algorithms of Association Rule, a series of emulation experiments show that this method has the effectiveness and feasibility as an efficacious attempt of refactoring XML structure.