Showing papers in "Computer Simulation in 2006"

Method for Selecting Parameters of Least Squares Support Vector Machines and Application

Abstract: Three step search method for selecting the parameters of least squares support vector machines is presented Theoretical analysis indicates that this method is effective and superior for selecting the optimized parameters of least squares support vector machines The steel’s harden-ability is predicted using this method for selecting the parameters of least squares support vector machines The simulation result shows that this method can get optimized parameters and precise modeling effect

Study of Automated PCNN System Based on Genetic Algorithm

Abstract: Pulse coupled neural network(PCNN)based on Eckhorn's model finds many applications in image processing.Because the parameters greatly affect the performance of PCNN,finding the optimal parameters becomes an onerous task..Especially in image segmentation,the parameters vary with the image that needs to process.An automated PCNN system was proposed that banded PCNN and genetic algorithm together and it was used to segment the image automatically and successfully.The correctness and dependability of the automated PCNN system are verified by experiment results,that is to say,the quality of the segmentation based on the automated PCNN system is much better and time-consuming and parameters-setting automatically is the main feature of the system.

Improved Adaptive Genetic Algorithm and its Application Research in Parameter Identification

Abstract: An improved adaptive genetic algorithm(IAGA) was proposed to avoid the premature problem and the slow convergence.Through the experiment of a typical Needle-in-a-haystack problem,the proposed algorithm shows its better global optimal ability and its faster convergence ability.Based on the above,the improved algorithm was applied to identify system parameter.The identification results show that this method has the advantages of high parameter identification precision,strong ability of resistance to the noise,good input signal generality and identification of the nonlinear system,so it has important practical values.

Adaptive Particle Swarm Optimization Algorithm and Simulation

Abstract: The reasons of premature convergence of Particle Swarm Optimization (PSO) Algorithm were analyzed. An adaptive PSO (APSO) was proposed, in which the inertia weight of the particle was adjusted adaptively based on the premature convergence degree of the swarm and the fitness of the particle. The diversity of inertia weight makes a compromise between the global convergence and convergence speed. Simulation results show that APSO not only effectively alleviates the problem of premature convergence, but also has fast convergence speed.

Public Opinion Propagation Model Based on Small World Networks

Abstract: Relationship network is a carrier of public opinion propagation. A relationship network topology based on small world model was built firstly. Then a basic computer simulation way of public opinion formation by designing individual’s local interaction rules and introducing the influence of individual’s psychology and mass media was proposed .At last, a public opinion spreading model in the special district of our country was designed and analyzed which provided a significant exploring method.

Selection of Kernel Functions and Parameters for Support Vector Machines in System Identification

Abstract: It is an important issue for support vector machines (SVMs) to choose kernel functions and parameters in engineering applications and theoretic research because SVMs with different kernel functions and parameters have greatly different performances. After the nonlinear system identification method using SVM was introduced briefly, the selection of kernel functions and their parameters was discussed in detail. Also, several SVMs with different kernel functions and parameters were used to identify nonlinear systems. A large number of experimental results show that the SVM using radial basis kernel function (RBKF) has good identification results than other kernel functions when SVM is applied to identify a nonlinear system. Furthermore, the parameter of RBKF is easy to choose. The space complexity is not large when the parameter of RBKF varies. So RBKF is a good choice for SVM in system identification.

Simulation Investigation on Rumor Spreading on Scale-free Network with Tunable Clustering

Abstract: This paper investigate rumor spreading on a triad scale-free network with power law degree distribution and tunable clustering.The objective is to study the impact of tunable clustering on the rumor spreading.Transformation of rumor spreading of different clustering is investigated by altering the particular parameters.From simulation results,a network with high clustering shows a good resistance to rumor spreading,while one with low clustering shows fragility to rumor spreading.

The Maximum Power Point Tracking Control of a Photovoltaic Power System

Abstract: Based on the simulation software,a simple and practical photovoltaic energy system is designed.By analyzing the physical model and electrical characteristic of solar cell,a photovoltaic system simulation model is introduced.The model is tested by using Matlab/simulink software.Considering the solar fluctuation and randomness,the model possesses the MPPT function.Also the article provides detail parameters for the design of the photovoltaic energy system.Simulation results demonstrate excellent tracking performance.

Subdivided Locating Method of Single Star Image for Star Sensor

Abstract: The characteristic of single star image was introduced. Some useful subdivision methods for improving the precision of star locating were expounded and analyzed. These methods were applied into star locating of simulative single star image. A detailed comparison was carried out for the precision of various subdivision methods when the star center of simulative single star image was located anywhere between the CCD pixels. The simulation results show that squared centroid and Gauss curve fitting methods are better, we can make a choice according to the condition of application.

Study on Optimization of SVR Parameters Selection Based on PSO

Abstract: The regression accuracy and generalization performance of the support vector regression (SVR) models depend on a proper setting of its parameters. An optimal selection approach of SVR parameters was put forward based on particle swarm optimization (PSO) algorithm. Furthermore, a comparison was made between the performance of PSO parameter selection and cross validation (CV) and leave-one-out (LOO) method on various data sets, such as a sin c function with additive noise and a SVR model of the product concentration in fermentation process. Simulation results show that the optimal selection approach based on PSO is available and the PSO-SVR model has superior learning accuracy and generalization performance.

Survey of Box-based Algorithms for Collision Detection

Abstract: Different algorithms for collision detections(CD)in virtual scene were surveyed.The principle and effectiveness of axis-aligned bounding boxes(AABB)method,oriented bounding box method(OBB),discrete orientation polytopes(k-DOP)method and Space-Time Bounding Boxes(STBB)method were discussed in detail.The results of analysis prove that the simpleness of bounding box is in contradiction with the compactness of wrapped object and the k-DOP method is between AABB method and OBB method in the effectiveness of CD.The main results are useful for designing virtual scene.

Distributed Peer-to-Peer Trust Model and Computer Simulation

Abstract: As the traditional concentrated trust method could not be used in peer-to-peer networks,a distributed peer-to-peer trust model was proposed.A peer's trust value was calculated according to the historical transaction information of the peer in this model.This model used an easy method to calculate the trust and avoided the divergence of iteration for trust calculation.In addition,this model could deal with some security problems,such as impostor attack and slander.The computer simulation results testify these merits.

Support Vector Machines For Time Series Regression and Prediction

Abstract: A method for predicting time series based on support vector machines was proposed. The time series, including simulated data and landslide deformation data sets, were preformed for regression and prediction by support vector machine, RBF networks, and elman recurrent neural networks. A comparison of these three methods was made based on their predicting ability. The results show that: when noise level is lower in simulated experiment, support vector machine is perfect relatively, and the Elman and RBF network are of more instability, on the other hand, with the higher noise levels, the greater relative error of two networks models is made. For landslide data sets prediction, the neural networks are limited to predict short term nonlinear time series in terms of their accuracy, whereas support vector machine has a higher precision in the short term and long term.

A Hybrid Particle Swarm Algorithm with Embedded Chaotic Search

Abstract: Evolutionary-based particle swarm algorithm has been investigated in this paper.A new hybrid particle swarm algorithm with embedded chaotic search is proposed for optimization in order to avoid trapping in local optimum and stagnancy of population.This hybrid method makes use of the ergodicity of chaotic search to improve the capability of precise search and keep the balance between the global search and the local search,which makes population escape the stagnancy rapidly.It has been compared with other methods such as standard particle swarm algorithm,standard genetic algorithm and improved particle swarm algorithm. In comparison,the proposed method shows its superiority in convergence property and robustness.It is validated by the simulation results.

New Method Used in Gene Expression Programming:GRCM

Abstract: Gene Expression Programming is a new genetic algorithm based on genome and phenome.This algorithm is very effective,and the experiments show that it outperforms GP more than two orders at convergent rate for many problems.A new Gene Read Compute Machine method based on Gene Expression was proposed.In this method,the chromosome's fitness could be computed directly without transforming the chromosome into expression tree.The experiment shows that this algorithm is not only simple and effective,but also improves the speed of computing.

The Methods for Moving Object Detection

Abstract: Moving object detection is the basis for object tracking, traffic surveillance and behavior analysis. Due to the impact caused by background, shadow , change of illumination, moving speed and so on, moving object detection fails easily. So it is import ant to detect moving object better. This article discusses how to use moving object detection widely, the problems and difficulties faced and the basic categories of approaches etc. main merits and shortages of all approaches are discussed, and the development trend of the area is forecasted.

VORONOI Diagram-based Path Planning for UAVs

Abstract: Path Planning for uninhabited air vehicles (UAVs) plays important roles in autonomous flight and completing mission. On the basis of analyzing characters of VORONOI diagram, VORONOI diagram-based path planning arithmetic was researched in the background of Scenario and Mission Planning Federate in UAVS Simulation Federation. How to build VORONOI diagram based on threaten locations, compute costs of flight routes, choose preliminary path and optimize path were discussed importantly. Last, the advantages and disadvantages of the arithmetic were summarized and the improvement approach was proposed.

Voronoi Algorithm Model and the Realization of Its Program

Abstract: The Voronoi polygon network is used usually as a computer simulation model in research flelds of natural and social sciences. It is defficult to achieve program design for the Voronoi definition. Researchers like to use Delauney triangle to build their computer program model. This paper introduces the principle of the Voronoi polygon and the Delauney triangle and their forming method first, then discusses their -program realization. These initial network models have 100 - 400 polygons and periodic boundary conditions so the boundary effects can be neglected, and finally the frame of Turbo C and C++ program is provided.

Modeling of Nonlinear Two-tank System and Model Predictive Control

Abstract: By using hydrodynamics theory,the mechanics principle of two-tank system was analyzed and the system nonlinear mathematics model was established.The parameters of the system model were gained from the metrical data,three curves of simulation and real-time proved the validity of system state equation and reasonableness of control method in special control area.Constrained one-step nonlinear model predictive control algorithm was proposed.And input constraints,input rate constraints and output constraints were considered.Experiments in the real tank system proved the correctness of the method and effectiveness of the proposed controller.

Improved Particle Swarm Optimization and its Application Research in Tuning of PID Parameters

Abstract: Aiming at the problem that the particle swarm optimization (PSO) is difficult to deal with premature and local convergence, an improved particle swarm optimization with mutation (IPSOM) was proposed. This new algorithm introduces mutation operator into the particle swarm optimization and replaces those particles flying out the solution space with new particles during the searching process to overcome the shortcoming of the particle swarm optimization. Through testing with a typical Rastigrin complex mathematics function, the experimental results show that the improved method not only has better ability to converge to the global optimum than the PSO and the simple genetic algorithm (SGA), but also can avoid the premature convergence effectively. Based on the above, this improved algorithm was applied to design the PID controller of a high-order system with time delay. The results show that the approach is effective and the designed controller has excellent performance.

Combination of Virtual Reality and Supervisory Control System in Substations Operator Training System

Abstract: An interactive substations operator training system with virtual reality technology,compared with traditional operator training systems,provides embedded senses to trainees and decreases on investments for hardware.More actual senses of real surrounding were realized through combination of scenes of virtual reality with a real-time supervisory control system for a substation,in order to serve for more advanced training requirements.An effective method was proposed to structure a substation simulation system of efficiently interactive operation.In this system,a supervisory control system was linked with scenes of virtual reality and both of them reacted to each other synchronously.Furthermore,the procedure of establishing a communicational program was described,which was associated with the model of substation made by MATLAB,connecting the scenes of virtual reality with the substation real-time supervisory control system according to IEC870-5-103 protocol.

Modelica and Its Mechanism of Multi-domain Unified Modeling and Simulation

Abstract: Modelica and its main characteristics were first introduced, and then the mechanism of multi-domain unified modeling and simulation with Modelica was proposed The reason that Modelica is suitable for modeling complex system was analyzed The method for complex product modeling with Modelica was discussed, and the current status of the development of Modelica based modeling and simulation tools was proposed Finally, the advantages of multi-domain unified modeling with Modelica were summarized

Parameter-Settings for the Dynamic Simulation Based on ADAMS

Abstract: During the dynamic simulation of ADAMS,different results can be achieved by different parameter-settings.Based on the process of the numerical calculation for dynamics,a comparison research among different integrators and their formulations is proceed,which tells the merits,demerits and the use-occasions of both.Also an introduction about the other integrator-settings for dynamic simulation is explained,the causes are given of the failures for the dynamic simulation are analyzed and some remedies for each failure.To testify the truth of the analysis,an actual example is specified.At last,the ecumenical approaches and principles of the parameter-settings for the dynamic simulation based on ADAMS are brought forward,which are the basis for the best simulation results.

Research on Complex Product Development Process Simulation and Its Optimization

Abstract: According to the characteristic of complex product development process(PDP),an integrated model,which combines and then extends the process models of activity networks and design structure matrix(DSM),was proposed.Based on this model,a simulation algorithm was given to predict and analysis the performance of PDP.An optimization method,which combines the simulation algorithm and generic algorithms(GA),was given to improve PDP by its architecture.

Soft Sensor Modeling Based on PCA and Support Vector Machines

Abstract: Soft sensor is an effective method to estimate variables which are difficult to be measured on-line in industrial processes.Support vector machine(SVM)is a novel machine learning method based on the statistical learning theory.A soft sensor based on principal component analysis(PCA)and Least Square SVM was proposed.The PCA method could not only solve the linear correlation of the input and compress data but also simply the SVM structure.Cross validation method was used to select parametrs of LS-SVM model.Soft sensor was applied to prediction of 4-CBA.Results indicates that this method features high learning speed,good approximation and well generalization ability.It provides convenice for on-line 4-CBA measurement.

Model of HVDC Control System Based on Hybrid Petri Net

Abstract: The hybrid characteristic of HVDC control system was described, which consists of continuous dynamic behaviors and discrete event behaviors. A powerful modeling tool for hybrid system - hybrid Petri Net (HPN) was introduced to model HVDC control system. The proposed model consists of discrete logic layer and continuous layer. The layers were connected by a state identification interface. The discrete logic layer switches HVDC control modes and coordinates regulators in the continuous dynamic layer. The continuous dynamic layer regulates the continuous state variable of HVDC system in real time. The interface identifies the system state and drives the state changes of discrete logic layer. Simulation results show that the proposed hybrid model responds well to normal and faults state of HVDC system.

Complete Orthogonal Function System V and Points cloud Fitting

Abstract: A novel approach to points cloud fitting was given.The work is based on a class of complete orthogonal k-degree polynomials in L2[0,1](called V-system).V-system which is composed of piecewise polynomials including every degree polynomials has advantageous properties of multi-resolution analysis and local.The expression of k-degree(k = 0,1,2,3) V-system was given,which effectively generated geometrical sculpt from surveying object displayed of points cloud by using limited number items of Fourier-V progression.The experimentations indicate that this approach has simple processing,fast calculating and high precision for some ordinary project.

Pensim Simulator and Its Application in Penicillin Fermentation Process

Abstract: A detailed mathematical model of penicillin fermentation process and a simulator named Pensim for this process were discussed. Then a descriptive flow chart and a detailed fundamental structure of Pensim were given. There were no hard restrictions on the input variables and the kinetic parameters in this simulator although ranges of various inputs and parameters were suggested. For this reason, the revisory ranges of various inputs and parameters were proposed after intensive research. The results show that the revised Pensim simulator is more reasonable and its outputs are more physically meaningful.

Modeling Method for Large-Scale Cultural Heritage Sites and Objects Using Real Geometric Data and Real Texture Data

Abstract: Digital preservation of cultural heritage sites and objects has become a global problem.A real 3D digital method for culture heritage sites and objects was proposed using 3D laser scanners and CCD cameras.Firstly,the laser scans were preprocessed for noise removal and holes filling.Next step was using an improved ICP algorithm to register multiple laser scans step-by-step.Finally,a volumetric-based algorithm was used to construct a coherent 3D mesh that enclosed all range scans,and real 3D models of cultural heritage sites and objects were created through texture mapping.The examples of constructing 3D models of buildings and grottos have been presented,and the prefect results have been obtained.

A Color Image-adptive Watermark Based on Wavelet Transform

Abstract: As an effective way for protecting copyright of digital media,watermark is receiving more and more attention.In this paper,a new method of digital watermark for still color image based on the wavelet transform is proposed.In this method,the watermark is a bilevel image.First,the raw image in multiresolution is decomposed using DWT,then the watermark image is permutated and added into the horizontal,vertical and diagonal wavelet coefficients of multiresolution representation of blue component of the original image in the L layer.Finally the image with watermark is reconstructed using IDWT.Experimental results show that the proposed method is robust against JPEG compression,image processing and attack such as filtering and noise,and receives a good tradeoff between con-cealment and robustness,and the detecting procedure does not need origin host image by "majority principles" and improves the dependability of detecting the watermark.