finding global minimum of function

I have a function like y=f(x,k,rho) that x is particle size, k and rho are constant parameters. we integrated this function over sample size distributions (like p(x), probability distribution of x) to find the average value of function over the distribution. also we measured this average value in the lab. now i want to use the integrated model inversely to find k and rho by minimizing the difference between measured average value and the calculated value of the model by changing k and rho. using excel solver each time I achieve different paires of k and rho depending on initial guesses. how can I find the global minimum using matlab?