Graph explorer

Munchausen Iteration

We present a method for solving polynomial equations over idempotent omega-continuous semirings. The idea is to iterate over the semiring of functions rather than the semiring of interest, and only evaluate when needed. The key operation is substitution. In the initial step, we compute a linear completion of the system of equations that exhaustively inserts the equations into one another. With functions as approximants, the following steps insert the current approximant into itself. Since the iteration improves its precision by substitution rather than computation we named it Munchausen, after the fictional baron that pulled himself out of a swamp by his own hair. The first result shows that an evaluation of the n-th Munchausen approximant coincides with the 2^n-th Newton approximant. Second, we show how to compute linear completions with standard techniques from automata theory. In particular, we are not bound to (but can use) the notion of differentials prominent in Newton iteration.

5 nodes4 linksoverview mapMunchausen Iteration
5 nodes4 links
Munchausen Iteration5 visible / 5 total nodes / 5 links
Co-authorshipAuthorshipAuthorshipTopic signalTopic signalWMunchausen Iterationpreprint / 2016ARoland MeyerResearcherASebastian MuskallaResearcherTData Structures and Alg...3564 worksTSymbolic Computation372 works
PaperSignal 104 links

Munchausen Iteration

preprint / 2016

Open