On Computing an Initial Efficient Extreme Point