The standard random generator can be used 'rand' This provides a  ' random' number between 0 and 1 with a uniform distribution. Further for Weibull e.g, de pdf is given by

F(t)=1-exp(-(t/a)^b). 

Inverting this equations gives

t(F)=a*(-log(1-F))^(1/b)

so generating a MXN matrix of random liftimes following this Weibull distribution can be simply realised by

F=rand(M,N);

t=a.*(-log(1-F)).^(1/b);

For the other distributions the inverse function is only in approximation available. Look in the literature or internet for these approximations.