I'm supposed make some reductions but don't even know where to start. Any help would or explanation on how to do this would be much appreciated.
A reduction here would mean showing that if you had an efficient algorithm A to solve the discrete logarithm problem, then you could use that to construct an efficient algorithm B solving the computational Diffie-Hellman problem. (For starters, assume A always works and construct B that always works. For full credit, given A that works with some probability [over a uniform input], construct a B that works with the same probability [over uniform inputs].)
If you understand what the discrete logarithm and computational Diffie-Hellman problems are, a reduction should be immediate.