How to write SVR's constraint from into a single function with loss function in linear case?

Question

The loss function is an e-insensitive loss meaning won't punish anything smaller than a setting e. The slide simply says if choosing the right $\lambda$, the SVR is equivalent to solve the right one.

However, I can't figure out their detailed relation.