Problem with dimension of a vector

Hi, my code is the this

rho = 0.03; 
c = 0.05;
eta = 0.2;
mu = 1.2;
sigma  = 0.8;
alpha = -rho/eta;
pzero=(eta*mu+rho*c)/(rho+eta);
Passo = 0.01;
pMax = 15;
N = 10;

dt = 1e-2;
time = 0:dt:4;
W = zeros(1,length(time));
x = zeros(N,length(time));
rng(1);

pStar = zeros(1,N);
tauStar = zeros(1,N);

front=fzero(@(p)(p-c)*(sqrt(2*eta)/sigma).*quad(@(t) t.^(-alpha).*exp(-(t.^2)/2+((p-mu)/sigma)*sqrt(2*eta).*t),0,15)/gamma(-alpha)-quad(@(t) t.^(-alpha-1).*exp(-(t.^2)/2+((p-mu)/sigma)*sqrt(2*eta).*t),0,15)/gamma(-alpha),pzero+0.11);
cont=1;

for p = 0:Passo:pMax
    if p < front
        v(cont)=(front-c)*(quad(@(t) t.^(-alpha-1).*exp(-(t.^2)/2+((p-mu)/sigma)*sqrt(2*eta).*t),0,15)/gamma(-alpha))/(quad(@(t) t.^(-alpha-1).*exp(-(t.^2)/2+((front-mu)/sigma)*sqrt(2*eta).*t),0,15)/gamma(-alpha));
    else
        v(cont)=p-c;
    end
    asc(cont)=p;

      for j = 1:N
          for i = 1:length(time)-1
              W(i+1) = W(i)+sqrt(dt)*exp(eta*time(i+1))*randn;
          end
          ex = exp(-eta*time);
          x(j,:) = p*eta+mu*(1-ex)+sigma*ex.*W;
      end

      for j = 1:N
          for i = 1:length(time)-1
              if x(j,i) >= front
                  pStar(1,j) = x(j,i);
                  tauStar(1,j) = time(1,i);
                  funVal(cont,j) = exp(-rho*tauStar(1,j))*(pStar(1,j)-c);
                  break
              end
          end
      end

      meanFunVal(cont) = sum(funVal(cont,:))/N;
      cont=cont+1;
  end

The error is:

Attempted to access funVal(7,:); index out of bounds because size(funVal)=[6,8].

Error in pstar (line 56)
    meanFunVal(cont) = sum(funVal(cont,:))/N;

But if I don't write meanFunVal(cont) = sum(funVal(cont,:))/N; everything is ok. How is possible?? Is a simple vector and a simple sum. Can you help me?

Answer 1

[Thorsten]

For cont == 7, the expression x(j,i) >= front is always false so

funVal(cont,j) = exp(-rho*tauStar(1,j))*(pStar(1,j)-c);

is never executed, i.e., funVal(7,:) does not exist.

Not that many entries in funVal are never set, i.e, they are set to zero. Taking the mean averages across all values, including the zeros. I'm not sure if this is what you want.

If you don't want to average over the values not set, but only over the non-zero values, you can pre-allocate funVal at the start of the loop

funVal = nan(1501,10);

and after the for loop, compute the mean as

meanFunVal = nanmean(funVal');