psoprocess.m

粒子群算法的详细实现以及在大肠杆菌代谢过程中参数优化的应用。

function [Result,OnLine,OffLine,MinMaxMeanAdapt]=PsoProcess(SwarmSize,ParticleSize,ParticleScope,InitFunc,StepFindFunc,AdaptFunc,IsStep,IsDraw,LoopCount,IsPlot)
%function description:an n loop entire PSO, returen this min and max
%fitness value
%[Result,OnLine,OffLine,MinMaxMeanAdapt]=PsoProcess(SwarmSize,ParticleSize,ParticleScope,InitFunc,StepFindFunc,AdaptFunc,IsStep,IsDraw,LoopCount,IsPlot)
%inputparameter：SwarmSize: swarm size
%inputparameter：ParticleSize：dimension of one particle
%inputparameter：ParticleScope；
%         ParticleScope format:
%           3-D ParticleScope format:
%                                   [x1Min,x1Max
%                                    x2Min,x2Max
%                                    x3Min,x3Max]
%
%input parameter:InitFunc:initialize PSO fucntion
%inputparameter:StepFindFunc:one_step update v,position
%inputparameter：AdaptFunc：fitness function
%inputparameter：IsStep；IsStep＝0，iteration continues，or stop。default
%continue
%inputparameter：IsDraw；IsDraw＝0，draw iterations,or dont draw，default dont
%draw
%inputparameter：LoopCount:iteration times, 100
%inputparameter：IsPlot
%                 IsPlot=1；
%
%return：Result best solution after iterations
%return：OnLine data online
%return：OffLine data offline
%return：MinMaxMeanAdapt mean fitness value
%
%usage:[Result,OnLine,OffLine,MinMaxMeanAdapt]=PsoProcess(SwarmSize,ParticleSize,ParticleScope,InitFunc,StepFindFunc,AdaptFunc,IsStep,IsDraw,LoopCount,IsPlot);


%faults control
if nargin<4
    error('fault number of input parameter')
end

[row,colum]=size(ParticleSize);
if row>1|colum>1
    error('fault of particle dimension, is a 1*1 data.');
end
[row,colum]=size(ParticleScope);
if row~=ParticleSize|colum~=2
    error('fualt particle dimension scope');
end

%set default value
if nargin<7
    IsPlot=1;
    LoopCount=100;
    IsStep=0;
    IsDraw=0;
end
if nargin<8
    IsPlot=1;
    IsDraw=0;
    LoopCount=100;
end
if nargin<9
    LoopCount=100;
    IsPlot=1;
end
if nargin<10
    IsPlot=1;
end

%control if give below-2-D particle searching process
if IsDraw~=0
    DrawObjGraphic(ParticleSize,ParticleScope,AdaptFunc);
end

%initialize swarm      
[ParSwarm,OptSwarm]=InitFunc(SwarmSize,ParticleSize,ParticleScope,AdaptFunc)

%draw initial swarm position on the best function
if IsDraw~=0
    if 1==ParticleSize
    for ParSwarmRow=1:SwarmSize
        plot([ParSwarm(ParSwarmRow,1),ParSwarm(ParSwarmRow,1)],[ParSwarm(ParSwarmRow,3),0],'r*-','markersize',8);
        text(ParSwarm(ParSwarmRow,1),ParSwarm(ParSwarmRow,3),num2str(ParSwarmRow));
    end
end

    if 2==ParticleSize
        for ParSwarmRow=1:SwarmSize
            stem3(ParSwarm(ParSwarmRow,1),ParSwarm(ParSwarmRow,2),ParSwarm(ParSwarmRow,5),'r.','markersize',8);
        end
    end
end
    
%pause catch pics
if IsStep~=0
    disp('begin iteration, press any key')
    pause
end

%begin update algorithm
for k=1:LoopCount
    %show iteration time:
    disp('----------------------------------------------------------')
    TempStr=sprintf('%g iterations',k);
    disp(TempStr);
    disp('----------------------------------------------------------')
    
    %transfer one iteration algorithm
     [ParSwarm,OptSwarm]=StepFindFunc(ParSwarm,OptSwarm,AdaptFunc,ParticleScope,0.95,0.4,LoopCount,k)
    
    %draw below-2-D particle's new position on the aim function
    if IsDraw~=0
        if 1==ParticleSize
            for ParSwarmRow=1:SwarmSize
                plot([ParSwarm(ParSwarmRow,1),ParSwarm(ParSwarmRow,1)],[ParSwarm(ParSwarmRow,3),0],'r*-','markersize',8);
                text(ParSwarm(ParSwarmRow,1),ParSwarm(ParSwarmRow,3),num2str(ParSwarmRow));
            end
        end

        if 2==ParticleSize
            for ParSwarmRow=1:SwarmSize
                stem3(ParSwarm(ParSwarmRow,1),ParSwarm(ParSwarmRow,2),ParSwarm(ParSwarmRow,5),'r.','markersize',8);
            end
        end
    end
    
    XResult=OptSwarm(SwarmSize+1,1:ParticleSize);
    YResult=AdaptFunc(XResult);    
    if IsStep~=0
        XResult=OptSwarm(SwarmSize+1,1:ParticleSize);
        YResult=AdaptFunc(XResult);    
        str=sprintf('%g iteration the best aim value%g',k,YResult);
        disp(str);
        disp('next iteration, press anykey');
        pause
    end
    
    %record every step mean fitness
    MeanAdapt(1,k)=mean(ParSwarm(:,2*ParticleSize+1));
end
%for loop end

%record min and max mean fitness
MinMaxMeanAdapt=[min(MeanAdapt),max(MeanAdapt)];
%calculate online and offline performance
for k=1:LoopCount
    OnLine(1,k)=sum(MeanAdapt(1,1:k))/k;
    OffLine(1,k)=max(MeanAdapt(1,1:k));
    %OffLine(1,k)=min(MeanAdapt(1,1:k));
end

for k=1:LoopCount
    OffLine(1,k)=sum(OffLine(1,1:k))/k;
end

%draw offline and online performance curve
if 1==IsPlot
    figure
    hold on
    title('offline performance curve')
    xlabel('iteration times');
    ylabel('offline performance');
    grid on
    plot(OffLine);

    figure
    hold on
    title('online performance curve')
    xlabel('iteration times');
    ylabel('online performance');
    grid on
    plot(OnLine);
end

%record this time iteration best solution
XResult=OptSwarm(SwarmSize+1,1:ParticleSize);
YResult=AdaptFunc(XResult);
Result=[XResult,YResult];