For a general nonlinear optimization problem with both inequality and equality constraints, a corollary on the optimization problem with a lp penalty function is studied and proved based on the sequential quadratic programming approach. In particular, the corollary defines the inequality condition for the search direction of the optimization problem. Moreover, it proves that the derived search direction is a descent direction of the penalty function in the original problem by determining the range of penalty parameter.

The research that focuses on the robustness for the estimation of unknown parameters in the linear model has been a hot topic of statistics. In this paper, the robustness of Bayes linear unbiased estimator of the regression coefficients in the linear model in terms of error covariance and prior distributions are discussed, and the necessary and sufficient conditions for Bayes linear unbiased estimator holding its optimality when error covariance varies, or prior distributions vary, or error covariance and prior distributions vary, are respectively obtained.

For the problem of corridor recognition for autonomous mobile robot in the process of navigation in the indoor environment, an improved corridor recognition algorithm is proposed. The algorithm uses information of depth image provided by kinect sensor, and the depth difference between consecutive points and least squares are used to fit a straight line to detect the existence of corridor outlet. Experimental results show that the algorithm can efficiently recognize corridor types while having a robust, and is still able to maintain a high recognition rate of corridor types in a complex indoor environment.

In the cross-organizational collaborative optimal decision-making, the parameters of optimization usually originate from different subject data. It is difficult to solve the global optimization problems due to the lack of a trusted third party. A method combining the random matrix transformation and encryption technology is put forward to solve the collaborative optimization of the horizontally partitioned data. This method also overcomes the potential destabilizing impact of structural problems and structural solutions when using disturbance or difference algorithm. On one hand, the security protocol can ensure the consistency of calculation results with privacy-preserving and centralized results. On the other side, it can prevent the potential inference attack. The study can be widely used in collaborative computing security issues of optimization decision problems among enterprise alliance or supply chain.

Cascading failures in complex networks is of great significance for the stability assessment of complex networks. Based on the classical linear load and capacity model, a model resisting cascading failures is built by detecting the community structure of the networks and adding secondary tolerance value to community boundary nodes. In the process of cascading failures, different strategies are used to attack IEEE standard grids, domestic real grids. The simulation results show that the model can raise networks stability with lower cost by adding secondary tolerance value to community boundary nodes, finding that the community boundary nodes have the dual functions of "firewall" and "tipping point". By simulating the two-tier coupled networks except for single networks, the new model is also found abler to resist cascading failures in interdependent networks, which indicates that the proposed model has some flexibility.

The online advertisement is the most important model of Internet profit. It makes the computing advertising a hot point of current research. For the real-time requirement and the high accuracy requirement in the online advertising system, the difference between the implementation and efficiency is compared based on the two popular platforms of Hadoop and Spark. And then a linear model is proposed combined with the technology of Spark to be applied in the advertising system. Finally the linear model is optimized from the aspect of numerical characteristics, iteration and step size. The testing result shows that the accuracy has a 5% to 10% increase after the optimization.

Using the bibliometric analysis， social network analysis and key words co-occurrence methods, 959 papers from 1995 to 2015 published in China National Knowledge Internet (CNKI) are analyzed about brand extension, and the basic situation of domestic brand extension, hotspots and frontier are explored by quantitative research， based on the research of the high-frequency group, highly cited papers, keyword co-occurrence, the time evolution of keywords. The study shows that the issued amount of domestic brand extension research had an overall upward trend, and then got down in recent years. The main areas of brand extension research focus on the basic theory of brand extension, brand extension, brand extension assessment. At present, the hotspots and main areas of domestic brand extension focus on the old brand activation, brand extension fit and consumer brand extension evaluation.

As a swarm intelligence, the Artificial Fish Swarm Algorithm(AFSA) has its weakness in solving the problem of Multi-objective Knapsack, such as blindness search, low speed of convergence and low accuracy in solution. Combining the global information of the artificial fish position with improving the moving strategy of artificial fish self-adapting, an improved AFSA is proposed. Simulation on multi-objective knapsack problem shows that the convergence rate as well as the accuracy in the non-dominated solutions which have been found out in the improved AFSA is superior to Genetic Algorithm and Particle Swarm Optimization.

Establishing category keywords is the key problem in text classification, which should be solved first. On the basis of the classification of text by using the category keywords and TF-IDF algorithm, an improved TF-IDF algorithm has been proposed to overcome the shortcomings of the vector space model, which cannot well adjust the weights. Firstly, category keyword library should be established, and the expansion and duplication be carried out. The weight of keywords in the document is modified by the addition of the length of the document, and the shortage of the original features of the entry class distinction ability is solved effectively. By using Bayesian classification method, combined with the experiments, the effectiveness of the algorithm is verified, and the accuracy of text classification improved.

Based on the analysis of the characteristics of skin color information and hand motion properties, a gesture segmentation and tracking algorithm on account of the color set, the motion set and the vague gesture set is constructed and realized on hardware. With this algorithm, the skin color set is established in color space while the motion set in video flow space. And through blurring operation joining tracking algorithm and the background realtime updating to overcome the interference of complex background and light factors, the vague gesture set is obtained. The whole algorithm is implemented on the field programmable gate array (FPGA), realizing gesture real-time segmentation and tracking using streamlining and parallel processing technology. The experimental result shows that this method has good accuracy, timeliness and robustness. On a 100 MHz system clock, the speed of processing an image of 640×480 resolution can be as fast as 100 frames / second.

The distortion of background harmonic voltage is inevitable in grid-connected system, which would result in serious low-order harmonics in the injected grid current. The low-order harmonics can be suppressed by increasing the switching frequency, while it also brings many high-order harmonics. Based on PR+HC-based current controller, the effects of the current controller parameters on steady-state error, phase margin and gain margin in LCL-type grid-connected inverter are investigated. A method is given to obtain the controller parameters’ region that meets the requirements, and the optimal parameters could be obtained on the basis of analysis in system stability. Simulation results verify the proposed control scheme and parameters design method.

Mechanics and fatigue characteristics are important factors of the ropes rowed sling with sockets which affect the safe operation of telescopic arm of lifting machinery. A new cavity with wedge structure is designed based on the selection of 6×7+IWS wire rope and the whole 3D model of the rigging developed. By importing the model to Ansys-workbench platform, the finite element model is established. The degree of freedom in three directions is imposed to rigging’s pin at one end and loading the 100 MPa mass force at the other end as the boundary condition of the model. It is revealed that the inner cone angle is important which affects the mechanical properties of the rigging. The rigging can bear the maximum stress at least 342.17 MPa with 16° inner cone angle. As the inner cone angle changing in a certain range, the total deformation of the rigging is from 0.025 mm to 0.03 mm. The fatigue life of the ordinary lay wire rope can meet the requirement of the engineering crane, and it can be the preferred object.

(Ba _0.8Sr _0.2) _1-1.5x Bi_xTiO₃ (x= 0.02，0.05，0.15，short for BSBT) ceramics were synthesized by the traditional solid state reaction method. Sintering was done at 1473 K for 2 h in air atmosphere. The structural, microstructure and dielectric properties of ceramics were investigated, in which a diffuse phase transition was exhibited in dielectric studies and a diffuse dielectric anomaly was observed in high temperature range. According to the fitting results of Curie-Weiss law, BSBT ceramics exhibited a stronger phase transition for x=0.05. Besides, grain boundary resistivity of BSBT ceramics was characterized by an impedance analysis and activation energy and the activation energy for conduction calculated. Polaron theory indicated that the relaxation in BSBT ceramics at high temperatures was associated with the hopping ions caused by oxygen vacancies. And it is suggested that the diffuse dielectric anomaly for all samples was due to a competition between the dielectric relaxation behavior and the variation of the relaxation strength.

The head, bones, and other raw waste of Tilapia are hydrolyzed by elastase, and response surface analysis method is used to optimize the enzymatic processes. The enzyme concentration, temperature, pH, solid to liquid ratio, time and other factors are discussed to analyze the effect of hydrolysis based on the degree of hydrolysis and peptide content. The results show that the optimum conditions of elastase hydrolysis: enzyme concentration 0.553 mg/g, temperature 41.94 ℃, time 3.31 h, pH natural, solid-liquid ratio 1: 5, the degree of hydrolysis and the peptide content is separately 18.09% and 0.90 mg /ml in this condition.

In this paper, through calcining the ellipsoid-like MnCO₃ precursors at 500°C, the ellipsoid-like ε-MnO₂ catalysts with an excellent property to decompose methylene blue can be obtained. ε-MnO₂is measured by SEM, XRD, UV-Vis and BET. The results show that the ellipsoid-like ε-MnO₂catalysts possess a high surface area of ca.88 m²/g. The obtained ellipsoid-like ε-MnO₂ materials are used as catalysts to decompose methylene blue in the present property  of H₂O₂.And the ellipsoid-like ε-MnO₂ can remove about 92% of methylene blue, which shows excellent property to remove methylene blue at room temperature. This shows that ellipsoid-like ε-MnO₂ is a highly efficient catalyst removal of methylene blue.

The production, damage and treatment of municipal sewage sludge in Guangzhou city are discussed in detail. In combination with the standardization of sludge treatment in China, 5 processing techniques are discussed. According to the principle of reducing cost, saving energy and reducing consumption, current combination technology of Guangzhou and the economic performance of the routes are compared. It is concluded that the combined technical route of sludge treatment and disposal in Guangzhou is “concentration+mechanical dehydration +thermal dry→off-site or in-site disposal”, and the unit investment cost of sludge drying is 1,900,000 yuan /tDS, the total cost of operation is 19,660 yuan.The technical route is expected to become the standardized route for Guangzhou municipal sludge treatment and disposal.