This is homework so perhaps I can try to give you a couple of hints without providing you with the final solution. To import your files automatically, you'll probably need to loop through a list of your files and open them with audioread. To find the delay, yes, xcorr, or finddelay or alignsignals from the Signal Processing Toolbox should do the trick. Now I'm not sure what "layer" means in this case, but it probably refers to mixing your files, in other words, it's a simple addition. And if your signals have different sizes, you may want to zero-pad some of them, e.g. with padarray. Now this will cause your signal to clip (go above + 1 or below - 1), so you want to resize your final signal between these two values. And finally the audiowrite function will let you write your signal as a sound file.