Quote from: BlackHatCoiner on Today at 09:29:24 AM

Quote from: HmmMAA on Today at 08:41:24 AM

Good luck with your LN trying to solve an NP hard problem

.

Transaction selection is NP-hard, but the software isn't trying to find the ideal fee; just one that is cheap enough. That's why they use variations of Dijkstra and A*, and not these algorithms per se. In LND, for instance, you can check yourself how optimized it is, comparably to a simple Dijkstra algorithm implementation: https://github.com/lightningnetwork/lnd/blob/master/routing/pathfind.go#L494-L504.

But, I don't understand why you focus on routing algorithm being NP-hard, and ignore mining which is NP-hard itself. The typical way to resolve this, is just to select a "good enouch solution" rather finding the ideal. This is why you can find cases of blocks which could include higher paying transactions instead. System works pretty fine after all.

I agree , but mining isn't based on rapspis to work . To find an optimised solution in an NP-hard problem computational power is needed . Do you think your rapspi will be able to handle a massive amount of paths and liquidity if/when LN is implemented ? And it's not me that says it won't work , it's their devs . But still people in here claim that it will work . Should i listen to the devs or members in here ? Except if doomad is a pseudonym of some LN dev , in that case i'd like to hear his opinion . If not then what he says is pure bull crap.

https://lists.linuxfoundation.org/pipermail/lightning-dev/2021-August/003203.html
https://lists.linuxfoundation.org/pipermail/lightning-dev/2023-October/004154.html
https://njump.me/nevent1qqsv5jyfh9gp2qnhruevrph39s2mspzfw3e9mx5avg0wenltrdjnnhgpp4mhxue69uhkummn9ekx7mqppamhxue69uhkummnw3ezumt0d5q3yamnwvaz7tm0venxx6rpd9hzuur4vgq3vamnwvaz7tmjv4kxz7fwdehhxarj9e3xzmnyqy28wumn8ghj7un9d3shjtnwdaehgu3wvfnsz9nhwden5te0wfjkccte9ec8y6tdv9kzumn9wsq3yamnwvaz7tmnv9k8g6tkddsjummjvupzp3wd64ehu3l4gfkfm63yxqfpztkk97a8k5683p5pvphhnadtj6p2xgpax5