examp.m

这是在网上下的一个东东

%
% Reinitialize everything.
%
clear
rand('seed',0);
randn('seed',0);
%
%  This script generates the example problem 
%
%
%  First, the v,D transport parameters.
%
params=[1.0, 1.0];   % [v,D]
%
%  Locations and times of samples.
%
xsample=5:5:300;
tsample=300;
nt=100;
tmax=300.;
tmin=0.;
%
% And the time points at which we'll estimate the source concentration
%
if (tmax == tsample)
	t=linspace(tmin,tsample,nt+1);
	t=t(1:nt);
else
	t=linspace(tmin,tmax,nt);
end
%
% Make t a column vector.
%
t=t';
%
%  Compute the G matrix.
%
g=getrep('kernel',t',xsample,tsample,params);
%
%  Compute the correct data.  
%
cin=cintrue(t);
cout=g*cin;

% set the random number generator to its initial state so the
% problem is repeatable;
randn('state',0);

% compute data with measurement error (d vector)
d=cout+0.005*randn(size(cout));

%
%  Now, setup the prior expected value, lower bounds, and upper bounds, etc.
%
upper=1.1*ones(nt,1);
lower=zeros(nt,1);
%
% Gaussian expected value function.
%
expvalue=zeros(nt,1);
expvalue=exp(-((t-150).^2)/(2*20^2));
%
%  Noise level in the form noise = [epsilon_a epsilon_m].  
%
noise=[0.005 0.0];
%
%  Finally, have the MRE routine solve the problem.
%
[x,lambda,beta,a,p5,p95]=mre(g,d,upper,lower,expvalue,noise)
%
%
%
figure(1);
clf;
bb=plot(t,cin,'k-',t,expvalue,'k--');
set(bb(1),'linewidth',1.0)
set(bb(2),'linewidth',1.0)
aa=gca;
set(aa,'linewidth',1.0,'fontsize',18);
axis([0 250 0 1.2]);
H=ylabel('C_{in}(t)');
set(H,'Fontsize',18);
H=xlabel('Time (days)');  
set(H,'Fontsize',18);
H=legend('True Model','Prior Mean');
set(H,'Fontsize',18);
set(H,'LineWidth',1.0);
figure(2);
cc=plot(t,x,'k-',t,p5,'k--',t,p95,'k--');
set(cc(1),'linewidth',1)
set(cc(2),'linewidth',1)
set(cc(3),'linewidth',1)
axis([0 250 0 1.2]);
aa1=gca;
set(aa1,'linewidth',2,'fontsize',18)
H=ylabel('C_{in}(t)');
set(H,'Fontsize',18);
H=xlabel('Time (days)');  
set(H,'Fontsize',18);
H=legend('MRE Model','90% Interval');
set(H,'Fontsize',18);
set(H,'LineWidth',1.0);