Michiel De Jong's Ledger Loops does decentralized loop clearing with loop finding and 2-phase commit and "chunked clearing" (as I understand) and is very good invention.
I have a project from 2012 (which I finished last year, while solving coordination for decentralized multihop payments, see
here) that as a "byproduct" or accidentally happens to
potentially clear multilateral barter loops, although I cannot say exactly how it behaves.
I do "voluntary taxes" (yes, actually voluntary) by people choosing to add tax to their IOU and this assigns the "conductance" of the IOU, and IOU at a node compete for tax being redistributed (a governance function, it does not mean you get the tax, as you tend to usually pass it on. It means more you are entrusted to control the redistribution flow).
The thing is the tax continues to hop from node to node until it reaches a person without income (without incoming IOU). This means it will loop if there is a loop, as people all by definition have incoming IOU... But, it also diverts from loop if there is branches. But, even if it diverts (say half diverts), the remaining half does another pass, and then diverts half, so that the full amount actually passed both through the loop and the branch.
The tax is guaranteed basic income to people with no income. It lifts everyone out of poverty, is the idea.
I cannot say how this behaves in an actual network with many loops and branches and loose ends. I assume it may relate to how much is loose ends and how much is loops, i.e., how many have income and how many are "free loaders" (in Resilience people are meant to be able to live off the basic income, but here the issue is to what extent the redistribution will be clearing loops).
If, by any chance, the redistribution does tend to clear loops, you get "chunked clearing"
without loop finding and
without 2-phase commit with "hash lock" along nodes in loop.
But until it is verified, this is a, well, wild guess partly, but grounded in a true possibility.
Since my main project is multi-hop, this possibility was interesting to me (very interesting actually) but not essential. But the question now is: could a stand-alone multilateral barter system use this to do decentralized loop clearing!? Such a system would then be extremely minimalistic, and include a decentralized guaranteed basic income system as well. Pretty cool possibility.
Anyone who is very good at simulating these things mentally or mathematically, feel free to mention if you think it would tend to clear loops significantly enough to be used as a main clearing mechanism, or be mostly just very minimal clearing.
Peace Johan