Digital Memcomputing with Frog Jumps
摘要
Inspired by the impressive jumping abilities of frogs, we explore the potential benefits of incorporating jumps into the dynamics of digital memcomputing solvers. For this purpose, digital memcomputing solvers with and without jumps are applied across a range of Boolean satisfiability problems. According to our results, jumps can reduce the solving times by up to a factor of four and modify the scaling exponents. Interestingly, the advantages of using jumps can be seen in cases where their size is so large that otherwise the continuous dynamics of voltage variables becomes almost binary. It is also found that the functional form of the distributions of times to solution depends on the problem’s difficulty.