Optimizing Analog Computation: A Smart Dual-Railing Approach for Efficient Compilation
Andrei Migunov (Drake University), Khalid Mohammed (Drake University), Garrett Provence (Drake University), Nicholas Haisler (Drake University)
Analog computation models such as General Purpose Analog Computers (GPACs) and Chemical Reaction Networks (CRNs) enable continuous and molecular-scale computation. We build on our prior work presented at CCSC 2024, where we are developing a Python-based compiler, inspired by the work of Huang and Huls (2022), that converts GPACs into Population Protocols (PPs). In this work, we focus on a key early stage transformation known as Dual-Railing, which ensures that a given GPAC system is CRN implementable. We further optimize the current method, by applying Tarjan’s algorithm to analyze variable dependencies. This allows for a reduction in the size of the output system, thus improving efficiency throughout the compiler’s pipeline. This research was supported in part by Department of Energy Office of Science Award DE-SC0024278.