1、Artificial Fish Swarm Algorithm,By Yao Pi and Dazhi Shang,Brief presentation,,It emulates the behaviours of fish,The principle and description of AFSA,Its a novel optimizing method,but it also need to be modified,We draw into competition and survival rules,Hot Tips,Whats AFSA?,How does AFSA work?,Wh
2、ats the advantages and disadvantages of AFSA?,How to modify AFSA?,,Whats AFSA?,AFSA is a new intelligent optimization algorithm based on fishs behaviors It uses the fish bottom behavior process and finds the global optimalThe Artificial Fish acts their behaviors guided by some strategies and usually
3、 stays at the place with plenty food,,The principle of AFSA,The position of the fish pond can be expressed with vector: X=(x1,x2,xn),where xi(i=1,2,3,n) is the variable to be searched for the optimal value;AF food consistence at present position can be represented by:Y=f(x),and Y is the object funct
4、ion the distance between the AF individuals can be expressed as : = - ;Visual represents the visual distance; is crowd factor and Step is the step length.Rondom (step) produces a random number in the range 0,step.,,How does AF work?,AF_follow,AF_prey,AF_swarm,Three behaviors of Fish-swarm,,The behav
5、iors of AF,AF_Prey: Let express the AF states at present and select a state in its visual randomly,if ,which means that the fellow center is not very crowd,forward a step to the fellow center,otherwise execute the prey behavior.,,The behaviors of AF,AF_Follow: Let express the AF states at present an
6、d find ,in which is the biggest value of its fellows in the near fileds ,if Y/ , which means that the fellow has high food consistence and the surrounding is not very crowd,go a step forward to the fellow ;otherwise, execute the prey behavior.,,Its a novel optimizing method,has a strong robustness,N
7、ovel,Low,Adaptive,It has a low precision and low speed of astringency,It has a good gobal astringency and easy to be programmed,The advantages and disadvantages,,How to modify AFSA,We draw into competition and survial rules as the disadvantages of AFSA,Modified,2.survival,petition,3.others,Atificial
8、 Neural Networks,,Whats ANN?,It is a neural network model of animal behavior characteristics, the distributed parallel algorithm for information processing model,signals,neurons,Behaviors,analyze,process,,Activationfunction,Brief graphic description:,May have a number of input signals and output sig
9、nalsInput signals can be the linear signals or non-linear signalsHidden layer may be more than one layer and every layer may have many neurons,Input layer,Hidden layer,Output layer,signals,Output function,output,The nature of ANN,,For example,we can get the nature of the unknown function by graphics
10、,To determine what kind of working condition or failure mode,We can use ANN for?,ClassifyIt can separate different types of things,Regress,Pattern Recognition,,Unknown function：,In this example,we use ANN to get the nature of the unknown function by graphics,An example of ANN,F = 20+x1.2-10*cos(2*pi
11、*x1)+x2.2-10*cos(2*pi*x2),Code：,interval=0.01;x1=-1.5:interval:1.5;x2=-1.5:interval:1.5;F = 20+x1.2-10*cos(2*pi*x1)+x2.2-10*cos(2*pi*x2);net=newrbe(x1;x2,F) ty=sim(net,x1;x2);figureplot3(x1,x2,F,g);figureplot3(x1,x2,ty,b);,Renderings：,,Wavelet Neural Network,WNN,ANN,In most of our ANN system, we use step function as activation function.,Wavelet analysis and the result of combining neural networks,In most WNN system, we use logistic sigmoid function as activation function.,Thank You!,
