With the background conditions，this paper gives an accurate model to describe the passenger’s evacuation in ships． Also
an improved ant colony algorithm is given by using a new heuristic function and convergence conditions． The new algorithm can increase
the influence of social factors and psychological factors in personnel evacuation． The simulation result shows that it can successfully
solve the problems and it is better than basic ant colony algorithm in both time and efficiency． It can provide a new idea in this kind of
problems．

The classical newsvendor model ( CNM) is often used to study the supply chain coordinating problem with perishable products
in single period，and only one order is permitted in the CNM，which is inappropriate in fact． However，for some perishable products
with the special consumed time，such as the moon － cake for Middle － Autumn Festival，etc． ，the manufacture for these items provide
a price discount for advanced booking in order to reduce inventory level． This paper considered a two － stage supply chain composed
of a risk － neutral supplier and multiple competing loss － averse retailers，and investigated the combined impact of the competition
and retailer’s loss aversion attitude on the decision － making behavior of retailer and the coordination of supply chain within the
advanced － booking discount contract． Based on game theory，it shows that in this supply chain game，there exists a unique symmetric
pure Nash equilibrium，and the optimal total order quantity increases as the degree of competition increases but decreases as the loss aversion
increases． Moreover it is found that，advance － booking discount contract can coordinate the supply chain． At last，the effectiveness
of advanced － booking discount Contract is verified in supply chain by a case study．

A analysis method of the problems of interconnected systems and the decentralized controller design method are presented in
this paper． That the topology of the interconnection is arbitrary and the exchanged information between subsystems is lost cause the unreliable
link． Firstly，the stochastic variable responsible for the situation of the lossy communication is regarded as a source of model uncertainty，
and the system is modeled in linear fractional transformation． Then the robust control theory is employed for system analysis，
and the largest probability of communication failure，which is tolerated by the interconnected systems keeping mean square stable，can
be obtained． Decentralized state feedback controllers are designed to ensure that the whole systems are mean square stable for a given
communication failure rate，based on the technique of linear matrix inequalities． Finally，an illustrative example of power network
which is a part of China southern power grid is presented finally to show the effectiveness of the proposed model and method． Simulation
results demonstrate that when the communication faults occur between the power components in the China southern power grid，using the
method described in this paper can make the network well designed and wide-area controlled and the local power network mean square
stable．

By using the decoupled double synchronous coordinate system software phase locked loop ( DDSRF-SPLL) ，it can detect the
unbalanced grid voltage of positive sequence component and negative sequence component，and put forward the coordinate transformation
of the mathematical to establish the grid side of directly-driven permanent wind generator's mathematical model，which includes
network side PWM inverter control and Voltage phase locked loop． A kind of current decoupling control strategy is proposed． It uses the
space vector pulse width modulation ( SVPWM) on AC output current control． This method can independently control the output current
of the active and reactive component，and can make the output current waveform of sine and grid voltage in phase with the same frequency．
Thus it can greatly reduce the harmonic pollution． It can eliminate the influence of unbalance voltage，and to achieve a precise
phase-locked output current and gird voltage three-phase grid voltage amplitude imbalance． In this paper，using Matlab /Simulink software
platform，the simulation proves the validity and feasibility of the control system when the grid stays in balance and unbalance． This
system improves reliability of wind power grid connection．

Since the disadvantages and limitation of anti-collision system with radar in aircraft ( ACSR-anti-collision system with radar
in aircraft) ，the terrain avoidance strategy was proposed，which adopted the virtual radar to substitute the airborne radar system to solve
the problem of the terrain anti-collision． According to characteristics of aircraft motion，uniform motion model，uniformly accelerated
motion model and the current statistic model were established，which are suitable for fast and accurate prediction． The principle of the
interacting multiple model and Kalman filtering were adopted to predict the aircraft flight conditions in real time which were used to
match the airborne terrain database to avoid crash． The prediction arithmetic was testified by simulation，and the results show that the
prediction time has important effect on deviation． But the warning level can be determined according to the forecast time． The research
demonstrated that terrain avoidance is much more advanced than ACSR，and it can replace ACSR．

In this paper，the technical limitations of minimizing resource losing model and maximizing threat model are analyzed via reviewing
the current research of weapon target assignment ( WTA) problem． To solve the non － real － time problem with the existing
methods，a dynamic WTA method is proposed which accords with the practical combat． Firstly，to make the allocation model being real
－ time and general，the dynamic input of target is introduced at random． Secondly，aiming at the dynamic random system with target inputs，
the state of rebirth time is analyzed，and the system is transformed into a Markov process by embedded Markov Chain． Then the
system state transition probability is resolved． Thirdly，the objective function is constructed，and the target assignment dynamic model is
established by state transition matrix． Finally，the solution of model is given，and its validity and real － time characteristics are proved．
The method proposed in this paper refines and complements current weapon target assignment models．

To the speciality of mixed noise constituted by pulse noise and Gauss noise，we present a comprehensive algorithm in this
text，which is based on the simplified PCNN model，utilizing several technique specialities of the model，selecting parameters properly，
and combining with mathematical morphology method，median filtering and wiener filtering． This method performs better than average
filters and median filters on hybrid noise reduction while retaining edges and detail information of the image． Experiments show that the
effect of eliminating grey image mixed noise which applying the simplified PCNN eliminating algorithm proposed in this paper is good．
This algorithm can show a big advantage when in the comparison with other algorithms． This algorithm not only can effectively filter hybrid
noise but also can excel in real-time tasks because of its reduced computation complexity． With the increase of the image populated
by blends noise，the advantage is obvious．

Using classical Lie symmetry theory，boundary control for a class of parabolic distributed parameter systems has been studied，
and the control laws on open loop and closed loop have been designed to control the system reaching the fixed stationary state． By
means of infinitesimal generators and invariance condition，the analytic control conditions on open loop and closed loop are presented after
having the classical Lie symmetry of the system，which is expressed by infinitesimal generators． The control purpose can be satisfied
under the two types of control laws by setting system parameters，initial condition and control purpose ahead． Comparatively，from the
simulation，the convergence of output error under open loop control is slower than that under closed loop control，and however there are
overshoot phenomena nearby the intake under closed loop control． The result proposed in the paper can be applied for controlling a class
of temperature or concentration models with conduction and convention characteristics．

Abstract: Wind power systems are often installed where is remote，inaccessible or the climate is not suitable for human to stay for long
time． So the traditional scheduled maintenance and breakdown maintenance can not meet the demands． It is very essential to ensure re-
liable and stable operation，and to reduce the cost of maintenance，condition monitoring and fault diagnosis of wind power system． The
existing fault diagnosis methods for the main failure components of wind power system，including gear boxes，generator，power electronic
devices，blade and so on，were introduced and classified． This work will provide an effective reference to the improvemet of the reliability of wind power system，reduction of cost and promoting the engineering process．

Aiming at solving the problem of poor availability of the physical parameters of a practical overhead crane，this thesis proposes
a parameter identification method based on energy reservation． It builds up regression equations for identification based on energy
reservation principle，and then applies non-negative least squares algorithm to deal with data，obtaining the controller design model of overhead
crane． This thesis has the identification experiments done on the overhead crane experimental system simulating the characteristics
of the operation of practical overhead crane and gets the mathematical model of the system． On the basis of the mathematical model，
an LQR controller has been designed to testify the accuracy of the identification results，confirming that the parameter identification
method can help a lot on control of overhead crane．

The voltage level to reactive power optimization dispatch and control problem is incorporated． A model of reactive power optimization
is established based on multi-objective differential evolution，which takes into account of loss minimization，voltage level best
target． Considering the drawbacks of traditional differential evolution ( DE) algorithm such as premature and slow search speed，a
strategy of self-adapting parameter improved differential evolution algorithm was proposed and first applied in reactive power optimization
problem． By adjusting the mutation F and crossover CR during the evolution process，the diversity of population is increased and the
global search area is expanded，which avoids algorithm into a local optimal solution，at the same time，the convergence speed is accelerated
later． The simulations are carried out on IEEE-14 bus system，and the results show the validity of the proposed algorithm．

On the basis of SVM( Support Vector Machine) multiclass classification，an improved DAGSVM ( Directed Acyclic Graph
Support Vector Machine) hand gesture recognition approach is put forward． Firstly，depth information of the scene is collected by using
Kinect sensor and hand region is obtained． Then feature vectors are extracted，which are used to train multiple binary SVM classifiers．
DAGSVM classifier is constructed using DAG topological structure with trained binary SVM classifiers and its structure sequence is improved．
Finally，the experimental results proved that the improved DAGSVM could reach higher recognition rate and can be used in the
control of intelligent wheelchair successfully．

A building block approach is presented to design Glareshield of a simulator． On the basis of its function categorization，this
approach aims at designing independent functional blocks through appropriately lowering relative dependence on connections between
blocks of operating board and control board，realising separation of function and structure，module segmentation and data communication
between modules． Furthermore，taking communication between blocks as an example，independent communication blocks were designed，
which effectively prevented frequent occurrence of system crashes due to code disorder in communication． Through experiments
in applications，it proves that the building block design approach can greatly improve the reliability of the simulator． Furthermore，this
approach could also be applied in the design of front roof and center control console．

he position and pose accuracy of the robot end depends on the geometrical parameters accuracy of each joint． An easy and practical method of parameter identification method is put forward to improve the pose accuracy． Using the robot kinematics equation depending on D-H algorithm，the relationship between the end and each joint is obtained． This particular method as well as the corresponding experiment is designed to solve the problems in the identification process，the actual geometric structure parameters are obtained according to the accurate measurement by laser tracker． The experimental results verify the effectiveness of the presented method．

Robust fault tolerant control of networked control system ( NCS) with variable sampling period and time delays was studied in
this paper． First，a discrete － time model with parameter uncertainties lying inside a polytypic framework whose vertices were determined
through the Real Jordan form approximation is proposed． Based on the model，an observer was proposed to estimate the fault and guarantee
the convergence speed of fault estimation． Furthermore，the robust active fault tolerant controller based on dynamic output feedback
was designed to guarantee the stability of the closed － loop system． A sufficient condition for the existence of controller to guarantee
the system asymptotically stable is established in terms of linear matrix inequalities ( LMIs) ． When the LMIs have an optimal solution，
explicit expression of the desired optimal H∞ tolerant fault controller can be determined． Finally，numerical simulation is presented to illustrate
the effectiveness of the proposed design techniques．

The grain of strip steel can be turned to uniform． And the work hardening and inner stress can be eliminated by continuous
annealing line． So it is the key for improving steel’s mechanical properties． But in a real production process，the mechanism of continuous
annealing is complicated． The coupling operation parameters could change the characteristics of strip steel，and the detection of steel
hardness has a long time lag，which brings grate obstacle to improve steel quality． PLS is used to build the relationship between average
trajectory process and hardness． It can guarantee the safety of production process by realizing hardness prediction and process monitoring
timely． Simulations in real process shows the feasibility and effectiveness of the proposed method．

As a key parameter in blast furnace smelting process，the temperature of molten iron is of importance for smooth operation of
blast furnace and the energy consumption． This paper studies on the important indicator for heat state of the blast furnace，namely molten
iron temperature． By taking advantages of both method of K-means clustering and support vector machine( SVM) ，a K-means clustering
－ based SVM model is proposed for predicting the temperature of molten iron． Firstly，the training sample data are divided into m
classes and m SVM regression prediction models are established accordingly． At the same time，a particle swarm optimization algorithm
is utilized to optimize the model parameters． Then，m discriminant functions are established to recognize which class the sample data
belongs to． Finally，the sample data are put into the corresponding class of regression model to predict temperature． Compared to the
standard SVM － based prediction method，the proposed method predict the molten iron temperature with a higher accuracy．

On-ramp metering is an effective tool for traffic management on freeway networks． In this paper，based on the ordinary differential equation( ODE) model of the ramp system，a new fuzzy self-tuning PI controller for freeway ramp metering is designed． By considering the upstream and downstream traffic flow information sufficiently，the designed PI controller can effectively suppress the influence of the system disturbances． Furthermore，a fuzzy logic is esigned to tune of the PI controller parameters automatically． Compared with both PID and ALINEA approaches，the simulation results illustrate the validity and efficiency of the presented method intensively．

To ensure virtual enterprise ( VE) reaches the desired goals，the owner should design effective mechanisms to avoid moral
hazard． In view of the VE exists principal agent relationship between the owner and the partner，we incorporate the partner’s fairness
preferences psychology and design an effective incentive mechanism considering the partner’s fairness concern to prevent the moral
hazard problem，and compared to the traditional principal agent model． Analysis results show that the fairness preference has changed
some of the conclusions of the traditional principal agent model that influences the effort level and the revenue sharing． The owner
should try to choose the partners with lower intensity fairness preferences psychology to avoid moral hazard in order to increase its own
profit．

The attitude control is the key of the automatic flight control for quadrotor UAV． In order to improve the robustness of quadrotor
UAV for various uncertainties，a robust adaptive backstepping controller is designed． The whole attitude kinematical model is obtained
and translated into a MIMO nonlinear system with generalized uncertainty． Because the design feature of the system is strict feedback，
the backstepping controller is designed． A robust adaptive function is designed to counteract the influence of the uncertainties，
which include external disturbance and interior parameters perturbation． The nonlinear tracking differentiator is introduced to estimate
the differential signal of the virtual control law and to reduce the“computer explosion”problem which is ubiquitous in the backtepping
controller． The closed － loop system is proved to be stable and converge exponentially through constructing Lyapunov function． Simulation
results are presented to corroborate the effectiveness and robustness of the proposed control strategy．

The construction and techno － realization of high － speed data acquisition system are presented． The FPGA is adopted to realize
the design of the hardware，which includes analog signals filtering plastic circuit，high － speed AD interface circuit，USB interface
circuit． The system application software design is also analyzed in detailed，which includes FPGA program design，USB firmware program，
upper computer application software，etc． The experimental results show the feasibility of the designed method．

On the basis of the analysis and comparison of the existing power cable insulation monitoring methods，it designs the online
monitoring system of the underground power cable insulation approach for the operational characteristics of the underground coalmine cable
lines． The system uses the low-frequency signal injection method，injecting the low-frequency signal into the cable through the reactor
neutral point． It constitutes a loop of the low － frequency signal，the cable insulation resistance and the Earth． Under the circumstances
of injecting certain low-frequency voltage，it can achieve on － line monitoring of the cable insulation，by detecting the value of
the low-frequency current-line，so the problems of poor reliability and low sensitivity on-line insulation monitoring of the power cable are
solved better． The experiment shows that this method has a low error and high accuracy in detection; it can monitor not only the single
cable insulation but also the multiple ones． It can apply to both the main cable circuits and the branch ones．

In view of the nonlinear characteristics of the SPMSM drive system，the unified PCHD modeling and speed control of PMSM
is presented based on the interconnection and damping assignment of energy-shaping method and port-controlled Hamiltonian systems
with dissipation( PCHD) theory，to obtain high performance of the speed control system． First，from an energy-balancing point of view，
uncertain system unified PCHD modeling is established including the inverter，SPMSM with iron losses and mechanical load． And then，
the passive controller of the SPMSM drive system is designed，and the nonlinear disturbance caused by the inverter is compensated using
the extended state observer． Finally，speed regulator is designed by using ADRC in order to get the desired q axis current，both the design
and the implementation of the controller becomes simpler and easier． The simulation results show that the proposed method can ensure
global stability，strong robustness． The system achieves a satisfactory dynamic and static performance．

对钢坯入库堆垛问题进行深入研究，建立了以最小化钢坯出库倒垛数为目标的数学模型，并设计了一种可以动态调整垛位顶层钢坯的堆放位置的DIBF(Dynamic and Improved Best Fit)算法。算法主要分为两个阶段，首先通过聚类算法将辊道上暂存钢坯形成入库批次；然后对入库批次通过DIBF算法进行批次指派垛位。通过钢厂实际生产数据对DIBF算法、IBF(Improved Best Fit)算法和传统手工计算方法进行验证。结果表明，相对于IBF算法和传统手工计算方式，DIBF算法不仅能够在限制可用垛位数的前提下减少倒垛次数，而且也能提高垛位的空间利用率，模型及算法可行、有效。

Time-delay is frequently encountered in many control problems． The existence of time-delay induces difficulty in stability analysis
and may thereby lead to oscillations or degraded performance in closed-loop systems． This paper investigates the stability analysis
of linear time-invariant( LTI) systems with time delay depending upon the solution of the characteristic equation． Not only the absolute
stability but also the relative stabilities the transient performances are depended on the locations of the characteristic equation roots． The
existence of time delay transforms the closed-loop characteristic equation into a transcendental equation with an infinite number of roots．
By investigating the real part and the imaginary part of the transcendental equation，the characteristic equation can be transformed into a
rotation equation with two vectors related to the system parameters． The characteristic roots on the given region boundaries can be determined
exactly and concisely． Numerical examples are provided to illustrate the proposed algorithm．

Aimed to improve the seam tracking precision while welding，this paper proposes a laser vision seam tracking system which
is composed of a laser vision sensor，a deviation correction controller，two stepping motors and a cross slider． While the tracking system
runs，the laser vision sensor measures position deviation between the seam gun’s current position and desired position，and the controller
determine the correcting value，and the motors drive the cross slider to modulate the seam gun’s position and posture so as to offset
seam tracking error． The prototype of the system was constructed． Seam tracking experiments show that the tracking error can limit to being
less than 0. 5mm． It is concluded the laser vision seam tracking system exists wide application prospect in precision welding engineering．

Aiming at the control problem of glide stage to cruise stage，considering rotation of the Earth and stamping constraint and initial
state constraint of the cruise stage，control law design model has been build． Optimal control problem have been translated to nonlinear
programming problem Using Gauss pseudospectral methods． Nonlinear programming problem has been solved by SNOPT． The simulation
results show that Gauss pseudospectral methods can solve optimal control law problems of glide stage to cruise stage．

水面无人艇系统存在有大惯性、长时滞、非线性、难以建立精确模型等问题，易受海浪等外界干扰的影响，采用传统的PID控制难以达到良好的航迹跟踪控制效果。为了实现更好的控制效果，基于广义预测理论，将其与PID控制相结合，设计了GPC-PID串级控制器，分别控制无人艇的转艏运动和操舵运动，并采用分离式控制方案，通过航向控制间接实现无人艇的航迹跟踪控制。从matlab仿真实验结果可以看出，GPC-PID控制器具有良好的航向控制与航迹跟踪控制效果，具有响应速度快，控制精度高，鲁棒性好，抗干扰能力强等优点。