We considered blowup phenomena of solutions to the initial boundary value problem for a class of Newtonian filtration equations with dissipative gradient and nonlocal source terms by energy estimation method. We gave the condition whether the blowup of solution occurred or not. Then by using the suitable auxiliary function and Sobolev inequalities, we estimated the upper and lower bounds of blowup time of solutions.

By using the asymptotic priori estimates and the method of operator decomposition, combined with the modified pullback attractor theory, we studied the longtime behaviors of solutions for abstract evolution equations with memory type under the propagation of timedependent speed, and obtained the existence and regularity of timedependent global attractors.

We considered a class of dynamic model of human immunodeficiency virus (HIV) transmission and proposed a class of HIV transmission dynamic system by using the rule of population transmission in epidemic contagion region. Firstly, using the functional homotopic mapping method, we obtained the sequence of approximate solutions of a class of HIV nonlinear transmission dynamic system and their  uniformly convergence, and gave an example to get all the approximate solutions. Then the quantitative and qualitative analysis of the solution for HIV transmission dynamic system was carried out. The results show that the expressions of approximate solution can regulate the distribution for the number of infected persons and susceptible persons.

The author considered a class of fractional elliptic equation with superlinear nonlinearity and nonlocal coefficient by using  critical point theory and variational method. Some results for existence of nontrivial solutions of this kind of problem were obtained without the AmbrosettiRabinowitz’s superlinear condition.

Let p,q be odd primes such that p>q, let G be a  group of order p³q³. By using the method of local analysis on finite groups, and by analyzing the different actions among the subgroups of G, we classified the group \%G completely, and got all the structures of the group.

Using the method and principle of fuzzy sets, the author gave by virtue pointwise and pointless depiction of colimit in the category of fuzzy modules. Firstly, by introducing the structural theorem of the coproduct in the category of fuzzy modules, the author obtained existence, uniqueness and structural theorem
 of colimit in the category of fuzzy modules. Secondly, the author constructed a constant system functor from the category of shapes of J to the category of fuzzy modules, and proved adjoint pair of a colimit functor and constant system functor. Finally, according to the adjoint isomorphisms relationship between the functor Hom and the tensor product functor , the author discussed the relationship between limits and colimits in the category of fuzzy modules.

By using the notion of fuzzy prenuclei mapping on fuzzy Quantale, we proved that the category of fuzzy Quantale was a reflective subcategory of the category of fuzzy Z-Quantale,  gave the concrete structure of the fuzzy Quantale generated by the image set of universal mapping, and discussed the fuzzy Galois connection constructed by the reflector.

We constructed Leibniz action and crossedmodule of Leibniz superalgebras, and gave the definition of Leibniz superpairs and non-Abelian tensor product of Leibniz superalgebras. Based on the structure of non\|Abelian  tensor product of Leibniz superalgebras, we obtained the related results of short exact sequence and homomorphism of Leibniz superalgebras.

By introducing independent parameters and intermediate variables, applying the weight functions and the technique of real analysis, we established a Hilbert’s integral inequality in the whole plane with a best  constant factor and a non-homogeneous kernel, and discussed its equivalent forms and some particular cases.

We considered the continuity of solution set mapping for two parametric generalized setvalued optimization problems in topological vector space. When ordered pair mapping consisting the objective function was lstrictly conequasiconvex, under the assumption of weaker constraint qualification, we obtained the optimization condtions for the continuity of the solution set mapping for two parametric generalized setvalued optimization problems.

In view of the numerical solution of American bond option pricing model, we constructed a fully implicit finite difference scheme and proved the stability of the scheme. Then, the modulusbased matrix splitting iteration methods were applied to solve the linear complementarity problems (LCP), and further compared it with the projected successive overrelaxation (PSOR) iteration method. Numerical experiments verified the effectiveness and robustness of the new methods.

We considered vertexdistinguishing Etotal coloring of complete bipartite graph K_6,n(6≤n≤38), and gave the vertexdistinguishing Etotal coloring of a class of special complete bipartite graph by using the methods of combinatorial analysis, reduction to absurdity and constructing concrete colorings. The results show that the vertexdistinguishing Etotal chromatic number of complete bipartite graph K_6,n is 5 when 6≤n≤10 and is 6 when 11≤n≤38.

Using the design idea of the structure of graphs and number theory, we applied the labelling to the study of graphical password, gave and proved the setordered graceful labelling, the strongly graceful labelling, properties of some vertices and the oddgraceful labelling which was formed by adding the leaves of each vertex.

By using the method of “deletionrestriction” in the arrangements, and by examining the ncycle and npath graphs, we gave the characteristic polynomials of the graphical arrangements corresponding to the wheel graphs and the join graphs of two paths.

Under the left truncated and right censored data, when the time variable T obeyed the generalized exponential distribution, in view of whether the scaling parameters were affected by covariates, we established two kinds of models, and used  the maximum likelihood estimation method to estimate parameters. The NewtonRaphson algorithm was used to solve the estimation of the parameters. The two models were applied to the transformer life data set and the Channing house data set, respectively, we obtained its survival function and hazard function.

We discussed the statistical properties of the almost unbiased estimate of Liu and the almost unbiased Stein ridgetype principal components estimate for the linear regression model under the balance loss function. We gave the risks of the almost unbiased Liu estimate  and the almost unbiased Stein ridgetype principal components estimate, under the balance loss function,  respectively,  and discussed the relationship between these two risks  under different conditions.

For  the  partially linear model with longitudinal data, we established the information criteria of variable selection based on empirical likelihood method, and proved the asymptotic properties of the new variable selection method. We simulated and compared the limited sample properties of parameter and information criteria based on empirical likelihood. The results show that the information criteria based on empirical likelihood method overcomes the difficulty that the parameter  likelihood functions are sometimes difficult to obtain, and the model selection effect is better.

Using the regularizing effect between the coefficient of the zero order term and the righthand side term of the equation, the author considered a class of elliptic equations with degenerate coercivity and lower order terms. When the righthand side term f of the equation was only in L¹, the author obtained a priori L^∞ estimate of its weak solution and proved the existence of its weak solutions.

By estimating the weighed momentum, we proved that the solutions to the compressible quantum NavierStokes equations on a onedimensional bounded domain would blowup in a finite time when a weighed functional associated with the initial momentum was large enough. The result shows that the equations have not such globalintime solutions when the initial momentum is large enough.

We proved the existence and uniqueness of distributionally almost automorphic solutions for some singular stochastic differential equations on a separable Hilbert space by using the block diagonal form of a linear operator and decomposition techniques of almost automorphic coefficients on the suitable space.

We considered a hybrid collocation method for solving terminal value problems of a class of fractional differential equations. Firstly, based on shooting method, we converted terminal value problems of fractional differential equations to initial value problems. Secondly, using the theoretical results of the initial value problems of fractional differential equations, we gave a hybrid collocation method for solving the terminal value problems. Finally, we verified the validity of the method to solve terminal value problems of fractional differential equations by numerical simulation.

Aiming at the problem that the partition of the item domain was not considered in the definition of regular association rules, we proposed that maximal association rules were useful complement to regular association rules. Using soft set theory and logical formulas on soft sets as the main tool, we solved the problem of the description of related core concepts for mining both regular and maximal association rules, and obtained unified mathematical characterization of these concepts. The example analysis shows that, compared with classical mining methods, maximal association rule mining method based on logical formulas on soft sets can effectively reduce redundancy and eliminate invalid rules, which improves the accuracy and mining efficiency of rule.

We proposed an improved least squares twin support vector machine (SMIILSTSVM) incremental learning algorithm based on ShermanMorrison theorem and iterative algorithm. It solved the problem that least squares twin support vector machine (LSTSVM) did not have structural risk minimization and sparsity. The experimental results show that the proposed algorithm has high classification accuracy and high efficiency, and is suitable for noisecontaining crosssample set classification.

In order to ensure the accuracy and diversity of the base classifier at the same time, we proposed an adaptive ensemble algorithm based on clustering and AdaBoost. Firstly, the training samples were divided into multiple clusters by clustering algorithm. Secondly, AdaBoost training was performed on each cluster to get a set of classifiers. Finally, these classifiers were combined according to the weighted voting strategy. The weights of each classifier were adaptive, we calculated the similarity between the test samples and each cluster and got the test samples’ classification confidence given by classifiers. The experimental results show that the algorithm can achieve a higher classification accuracy compared with representative ensemble algorithms such as AdaBoost, Bagging (bootstrap aggregating) and Random Forest.

Based on the uncertain data representation model, we proposed an uncertain data generation algorithm ACUDGen (attribute level continuous uncertain data set generation algorithm) for attribute level uncertain data. By introducing the outlier detection algorithmLOF (local outlier factor) algorithm, the algorithm used the outlier factor of each data object as the parameter to control the perturbation range of uncertain data objects, which could well satisfy the distribution characteristics of the original data and solve the problem of lack of the distribution characteristics of the original data in the present work. The experimental results show that the uncertain data set generated by the proposed algorithm has a better clustering effect, and reduces the influence of outier on the clustering results, so that the size of each data object MBR (minimum bounding rectangle) can be adaptively changed according to its own distribution characteristics.

Aiming at the characteristics of sparseness and too limited words in one short text, in order to better deal with the problem of short text classification, we proposed a short text classification model based on integrated neural networks. Firstly, the extended word vector was used as the input of the model, so that the numerical word vector could effectively describe the morphological, syntactic and semantic features of short text. Secondly, the recurrent neural network (RNN) was used to model the semantics of short text, capture the dependency of internal structure of short text. Finally, we used the regularization term to select the model with minimal empirical risk and model complexity simultaneously in the process of training model. By the short text classification experiments on the corpus, we verified  that the proposed model has a better classification effect, and the classification model could deal with short text input with variable length, and has a good robustness.

Aiming at the problems of the current attributebased encryption system, such as the issues of users and center abused keys and decryption rights of user revocation, we proposed an attributebased encryption scheme that supported accountability and user revocation. The scheme used a “fixed point” as the identifier of the key owner to realize the identity tracing of malicious users. The decryption key was generated by interaction between user and center to solve the problem that semitrusted center abused keys. By using complete subset scheme and selfupdate encryption scheme, the proposed system could achieve immediate revocation and forward/backward security. Furthermore, we used a Paillier encryption system to avoid storage cost in the tracing process. The scheme didn’t need to predetermine the number of attributes at stage of system setup, and was proved to be chosenplaintext security in the standard model.

Aiming at the deficiency of traditional B spline interpolation method in unmanned aerial vehicle route optimization path, we proposed a new fast route optimization algorithm based on the Voronoi diagram. Firstly, the initial path was generated in the Voronoi diagram according to the shortest path algorithm, and then the optimization algorithm was used to update and replace each node that made up the nonflying angle in the path to generate an optimal path. The algorithm solved the problem that the unmanned aerial vehicle could not fly because of the unreasonable initial path, and ensured the optimal cost of the threat and fuel cost in the flight path. The experimental results show that the algorithm satisfies the validity and rationality of route planning.

Aiming at the problem that it was not easy to deduce the transfer function of the controlled object when three parameters were adjusted in the traditional PID (prop
orationalintegraldifferential) controller, and these parameters were difficult to be adjusted manually, we proposed a new algorithm to adjust the PID controller parameters. The algorithm combined neural network with genetic algorithm. First, the algorithm used the simulation function of the neural network to assist the genetic algorithm calculating the fitness, and the trained neural network was used to simulate the controlled object. Then three parameters of the PID controller were constantly optimized in the evolution of genetic algorithm. Compared to the traditional parameter tuning, the simulation results show that the proposed algorithm has strong robustness and fast response speed.

Aimimg at the problem that the active contour model was difficult to achieve high precision segmentation of the image, in order to obtain more ideal image segmentation results, the author proposed a new image segmentation method based on improved particle swarm optimization (PSO) algorithm. Firstly, the traditional active contour model was analyzed, and the limitation of its existence was pointed out. Secondly, the objective function of the energy minimization control point was established, the optimal value of the objective function was searched by the particle swarm optimization algorithm, and the image segmentation was realized according to all the energy minimization control points. Finally, the standard image database and the traditional image segmentation method were uesd for comparative test. The test results show that, compared with the traditional method, the proposed method can segment the images more accurately and quickly. It can effectively suppress the noise interference in the images, and obtain ideal image segmentation effect.

Aiming at the problem that there were a large number of films, and it was difficult for audiences to choose their favorite films, we proposed a personalized recommendation method of film based on reviews. Firstly, the method classified the audiences’ reviews by using the film labels as the topics of film. Secondly, we calculated the satisfaction degree of topics according to the evolution function of the topic. Thirdly, we set the weights of topics according to audiences’ preference and obtained the personalized evaluation of the film. Finally, according to the evaluation results, audiences were provided with an alternative sequence of films. The experimental results show that the analysis method of film based on reviews is effective.

Aiming at the problem of low efficiency of parameter extraction in theoretical chemistry, with the help of list generation technique and thread simulation technique in Python, we proposed a Wfn^auto algorithm and used the proposed algorithm to test some typical wfn files. The test results show that the Wfnauto algorithm can significantly improve the efficiency of parameter extraction in analytical chemistry.

We simulated and calculated the radiation damage by proton in CdZnTe. Firstly, the ionization damage and displacement damage for protons in CdZnTe were calculated. Secondly, the number of vacancies generated by protons radiation at different energy and incident angles in CdZnTe were calculated. Finally, sputtering yields generated by protons radiation at different energy and incident angles in CdZnTe target with different thicknesses were calculated. The results show that ionization damage is much greater than displacement damage; the number of vacancies generated by proton radiation in CdZnTe increases with the increase
of proton energy, and the number of vacancies decreases quickly when the incident angles of protons are greater than 60°. CdZnTe is lower than radiation resistance of silicon and diamond semiconductor materials. The sputtering yield is much lower than the number of vacancies in CdZnTe, sputtering yield
increases first and then decreases with the increase of proton energy, and increases with the increase of incident angles of protons. The overall trend of sputtering yield increases with the increase of thickness of CdZnTe target.

The ptype coaxial highpurity germanium detector was used to measure the decay γ energy spectrum of ^226Ra which came from fluorescent instrument panel and ¹³³Ba standard source, the energy and intensity of characteristic γrays were analyzed. The results show that the characteristic rays of ^226Ra and a large number of characteristic γ rays of the decay daughter nucleus of ²²⁶Ra appear in the ²²⁶Ra spectrum. There are large number of sum peaks in ¹³³Ba γ spectrum except its own characteristic rays. The existence of sum peaks is proved from the relationship between counting rate and distance and γ ray cascade.

Based on the three methods of Mulliken, Stoutpolitzer and SCPA, we calculated the contribution of basis function of the frontier molecular orbital components of alanine (Ala) chiral enantiomers, respectively. The results show that three methods are used to calculate the contribution of basis function of the 20th molecular orbital component, the 5th and 9th basis functions, the qualitative conclusion that describes the maximum contribution of the basis functions describing the 2pz atom orbits of 1N atom is consistent. The 2p atomic orbital shells of each atom have limited contribution to the molecular orbits near the nucleus. The contribution to the 15—20 occupying orbit is more complicated, the results of 2p atomic orbital shell contribution of 1N,3C,5C atoms are consistent, the contribution of the 2p atomic orbital shell of 6C and 10O atoms is quite different.

The Hf_15-xCo₇₈B₇W_x (x=0,1,2,3) alloy meltspun ribbons were prepared through the method of adding trace of W in Hf₁₅Co₇₈B₇ alloy. The magnetic properties, phase composition and microstructure were investigated by magnetic measurements, Xray diffraction (XRD), thermomagnetic analysis (TMA) and scanning electron microscope (SEM). The results show that proper addition of W can improve the coercivity （H_C） of Hf₁₅Co₇₈B₇  alloy significantly, 
 Hf_15-xCo₇₈B₇W₂  meltspun ribbons are composed of Co7Hf phase and a small amount of fccCo phase, W atoms enter the lattice of Co₇Hf phase, resulting in the decrease of the Curie temperature (T_C) of Co₇Hf phase and increase of the magnetocrystalline anisotropy field (H_a). The grain size of alloy decreases significantly.

Based on density functional theory, we calculated the modulation of the electronic structure of vacancy defects of different concentration points on the armchair and zigzag carbon nanotubes, and analyzed the bond length, the defect formation energy, the bandgap and the electron state density. The results show that with the increase of concentration of defects, the bond length of the neighboring carbon atoms decreases, the bandgap of the armchair carbon nanotube opens, the metallic properties change to the semiconductor properties, and the bandgap of the zigzag carbon nanotubes gradually decreases. The absence of carbon atoms causes the nonbonding unpaired electrons near the defects to form additional electronic state near Fermi energy levels, which changes the electronic structure of the carbon nanotubes.

Using oxide graphene as sorbent, we established a solidphase extraction method for the extraction of five pyrethroid pesticide residues of cyfluthrin, tefluthrin, fenvalerate, permethrin and bifenthrin from ginseng  beverage. The extraction conditions were optimized by single factor experiment and orthogonal experiment. We investigated the effect of the amount of oxide graphene, the type and volume of eluent and the extraction time on the extraction. The results show that the enrichment effect of pyrethriod pesticide residues is better, the recovery rates of cyfluthrin, tefluthrin, fenvalerate, permethrin and bifenthrin  are 105.98%,65.93,105.17%,93.69% and 89.46% respectively under the optimal extraction conditions.

O-quaternized thiourea oligosaccharides was prepared with chitosan oligosaccharides as raw materials, and its structure was characterized by infrared spectrum (IR), UVVis absorption spectra and thermogravimetric analysisdifferential thermal analysis (TG-DTA) methods. Using the herring sperm DNA as a target, the action mode of O-quaternized thiourea oligosaccharides and herring sperm DNA was studied by UV-Vis absorption spectra, cyclic voltammetry (CV) and viscosity method. The experimental results show that the absorption peak of Oquaternized thiourea oligosaccharides appears at 213.5 nm and 280.0 nm, its thermal stability is poor, and the oxygenolysis begins at 180 ℃. The UVVis absorption peaks of Oquaternized thiourea oligosaccharides change after adding DNA, O-quaternized thiourea oligosaccharides reduce the current of the redox peak of the potassium ferricyanideDNA system, potential shifts positively, and the viscosity of DNA increases. The action mode of O-quaternized thiourea oligosaccharides and DNA is intercalation.

The isothermal multipleselfmatchinginitiated amplification (IMSA) technology was used to detect human papillomavirus (HPV) and six specific primers were designed for HPV16E7 or HPV52E6 gene sequences. Meanwhile, dual fluorescence indicator (hydroxy naphthol blue (HNB) and SYBR Green Ⅰ) was added to the detection system. We established dual fluorescence IMSA method for rapid detection of human papillomavirus. The results show that the dual fluorescence indicator constructed by 340 μmol/L HNB and 1∶10 000 SYBR Green Ⅰ has obvious indication effect in the IMSA reaction system. The coloration of positive reaction tube is yellowgreen under the 455 nm blue excitation,  the coloration of negative reaction tube is orangered, and the detection limits of HPV16 and HPV52 are 60,600 copies/μL, respectively, HPV16 and HPV52 can be specifically detected in the samples, and there is no difference in comparison with clinical detection results.

The PDL1 monoclonal antibody was transiently expressed by using HEK293 6E cells. We cultivated HEK293 6E cells in serumfree suspension and selected the optimum growth medium, investigated the transfection conditions of HEK293 6E cells, including culture medium varieties, the mass ratio of DNA and PEI, and cell density while transfecting, then identified the harvested antibodies by SDSPAGE and ELISA. The results show that the best transfection conditions were as follows: A style, a serumfree culture medium, a cell density of 2×106 cells/mL and m(DNA)∶m(PEI)=1. The highest yield of antibody was 170.9 μg/mL at the fourth day after transfection, the molecular weight of the heavy chain of the harvested PDL1 antibody was about 50 000 and that of the light chain was about 25 000.

Mung bean polypeptide was prepared by hydrolysis of papain and alcalase. First, the singlefactor experiments were carried out on the pH value, the mass ratio of alcalase and isolated mung been protein, the temperature of hydrolysis and the mass fraction of isolated mung bean protein of the hydrolytic system. Then the orthogonal experimental design with four factors and three levels was used to optimize the hydrolysis conditions of alcalase on the basis of singlefactor experiment. The optimal hydrolysis conditions were as follows: the mass fraction of isolated mung bean protein was 8%, the temperature of hydrolysis was 55 ℃, the ratio of enzyme to substrate was 8%, and pH=90. Under the optimal hydrolysis conditions, the degree of hydrolysis of isolated mung bean protein reached 3258%.

Caenorhabditis elegans was used as model to evaluate the antioxidant activities of five kinds of natural food and medicine homologous plants (Rosa roxburghii, Zingiber oj\|jicinale Rosc., Cinnamomum cassia Presl, Dioscoreae rhizoma, grape seeds) by C.elegans lifespan test, oxidative stress test in vitro and free radicals content test in vivo. The results show that 2 mg/mL of R.roxburghii and grape seeds water extractions both can prolong the lifespan of testedC.elegans  remarkably. Five kinds of plant water extractions with different mass concentrations have remarkable repair abilities of oxidative damage and certain free radical scavenging abilities to the tested C.elegans. Among them, the free radical scavenging effects of R.roxburghii and grape seeds water extractions are obvious.