Robust Signal Restoration in Chemical Reaction Networks

Molecular computing systems that are contained in well-mixed volumes are often modeled using chemical reaction networks. In these systems, concentrations of molecules are treated as signals and used for both communication and memory storage. A common design challenge for such a system is to avoid memory corruption caused by noise in the input signals. In this paper, we analyze two signal restoration algorithms for molecular systems modeled with chemical reaction networks. These algorithms are designed to prevent a memory signal from degrading over time, and we show that under modest conditions these algorithms will maintain the memory indefinitely. We also present an exact solution of the running time of the first algorithm which demonstrates that it converges in logarithmic time.