I will give this a crack as it appears unanswered and I have dealt with this problem in the past. The OP is correct, it is in fact possible for a rarefaction wave to overtake a traveling shock wave from behind. The classical problem usually arises from the one-dimensional motion of a piston inside a channel. The instantaneous motion of the piston sends a shock wave through an initially stationary gas. At a later time, the piston abruptly comes to a stop. A simple rarefaction wave is sent into the post-shock gas and eventually catches up to the traveling shock. Below is a schematic of the problem in the space-time ($x-t$) diagram.
Once the rarefaction overtakes the shock, a non-uniform region behind the shock begins to develop with entropy variations, which are carried along the particle paths in the region behind the shock wave. The entropy variations are what make the analytical approach very difficult. The governing equations for this system are,
$$ \frac{\partial \rho}{\partial t} + u\frac{\partial \rho}{\partial x} + \rho \frac{\partial u}{\partial x} = 0 $$
$$ \frac{\partial u}{\partial t} + u\frac{\partial u}{\partial x} + \frac{1}{\rho} \frac{\partial p}{\partial x} = 0 $$
$$ \frac{\partial s}{\partial t} + u\frac{\partial s}{\partial x} = 0 $$
The first attempt at a solution was carried out by Friedrichs 1948. His solution was for an infinitely weak shock where the entropy variations across the shock could be neglected. This assumed the flow behind the decaying shock was isentropic and the Riemann invariant is constant through the shock.
Other researchers provided ways to approximately account for the variation in entropy behind the decaying shock. Seek the works of Pillow 1949 and Lighthill 1950.
A solution for a moderately strong shock was presented by Ardavan-Rhad 1970. However, this solution only approximately satisfies the Rankine-Hugoniot shock jump conditions, and displays increasing error with high shock Mach numbers.
Finally, the solution you are seeking was provided by Sharma, Ram, and Sachdev 1987. The solution is for an arbitrary strength shock and describes the decay of the shock propagation from the interaction of an overtaking rarefaction wave. Moreover, the solution is uniformly valid for the entire flowfield behind the decaying shock wave. It is noted, the solution satisfies the governing equations exactly, but only approximately satisfies two of the Ranking-Hugoniot shock jump conditions. However, the error incurred is shown to be small for strong shocks. Also, in the limit of a weak shock, the solution converges to that proposed by Friedrichs.
I will be honest, the solution is mathematically heavy for myself. I have implemented the solution in Matlab for various calculations and have experienced some difficulty. Occasionally, I get solutions that are non-physical as their proposed solution method requires implementing a Newton-Raphson method very close to a singularity for a certain parameter. However, the more experienced may not have this issue. Anyways, I hope this answer/solution helps.
