Answer to Question #11699 in Quantitative Methods for Mark

Question

I have an equation BG(T) = BG(t) + AxC(t) - BxD(t), where BG(t), C(t) and D(t) are known at time t, A and B are constants and I am trying to forecast BG(T). A and B are not known with 100 % accuracy. Example: A is 3.0 plus or minus 0.05 and B is 2.0 plus or minus 0.05. Given the confidence interval in A and the confidence interval in B, how would I calculate the confidence interval of BG(T) ? If possible can I have an answer assuming the constants are i) independent and ii) not independent of each other. Thanks, Mark