MATLAB: Maximize Expected Utility with 2 Random Variables - Portfolio Optimization

I want to maximize a utility function for a portfolio optimization. The utility function has two decision variables a_1,a_2, which represent the ratio of the investment of the initial wealth between a risk-free investment and a risky investment. The interest rate of the risk-free option is known, but the interest rates of the risky option are normally distributed with given mean and variance-covariance, thus random.

I approximate the expectation with a Gauss-Hermite quadrature using the Miranda and Fackler Toolbox (http://ift.tt/1O2v7qL) and especially the function qnwnorm.

How do I get to include the decision variables a_1,a_2 to the qnwnorm in MATLAB, so that I can optimize afterwards the expectation of the utility function? I'm a newbie in these applications, so bare with me!