You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The best strategy seems to be his own Branch and Bound strategy, that tries to look for exact match. It's defined in the thesis on pages 35 and forward.
The code still a little ugly, but it seems to be working and seems to produce the lowest fees in the stats.
I will clean-up the code now and then think about how to do the tests, when the algorithm is probabilistic. (It has a random utxo selection as a fallback.)
edit: OK there is an error in the algorithm; when I corrected it, the fees are similar to the other ones in the simulation.
@xekyo wrote a master thesis, where he evaluated several coin selection strategies on a large dataset
http://murch.one/wp-content/uploads/2016/11/erhardt2016coinselection.pdf
The best strategy seems to be his own Branch and Bound strategy, that tries to look for exact match. It's defined in the thesis on pages 35 and forward.
The code seems to be here -https://github.com/Xekyo/CoinSelectionSimulator/blob/master/src/main/scala/one/murch/bitcoin/coinselection/StackEfficientTailRecursiveBnB.scala - MIT license
I am now trying to port the algorithm to javascript and to this module style.
See also bitcoin/bitcoin#7664 (comment) , murchandamus/CoinSelectionSimulator#3
edit: And also see c++ implementation here bitcoin/bitcoin#10637
The text was updated successfully, but these errors were encountered: