When disallowing error, traditional chemical reaction networks (CRNs) are very limited in computational power: Angluin et al. and Chen et al. showed that only semilinear predicates and functions are stably computable by CRNs. Qian et al. and others have shown that polymer-supplemented CRNs (psCRNs) are capable of Turing-universal computation. However, their model requires that inputs are pre-loaded on the polymers, in contrast with the traditional convention that inputs are represented by counts of molecules in solution. Here, we show that psCRNs can stably simulate Turing-universal computations even with solution-based inputs. However, such simulations use a unique “leader†polymer per input type and thus involve many slow bottleneck reactions. We further refine the polymer-supplemented CRN model to allow for anonymous polymers, that is, multiple functionally-identical copies of a polymer, and provide an illustrative example of how bottleneck reactions can be avoided in this new model.
CITATION STYLE
Tai, A., & Condon, A. (2019). Error-Free Stable Computation with Polymer-Supplemented Chemical Reaction Networks. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 11648 LNCS, pp. 197–218). Springer Verlag. https://doi.org/10.1007/978-3-030-26807-7_11
Mendeley helps you to discover research relevant for your work.