[Cbc] Badly scaled model: "Problem proven infeasible"

[rhavar at protonmail.com]

I'm working with a real problem [1], and running into it cbc saying "Problem proven infeasible" however there exist a few trivial solutions. I'm pretty sure the cause because it's "badly scaled" however I am really just doing a direct translation of the problem into a linear model. It's not really obvious for me how to solve it, the problem domain inherently has huge amounts of scale  (I'm working with bitcoin, where some amounts are 100e8 satoshis (100 BTC) but yet everything needs to add up to 1 satoshi) :/

Any guidance would be appreciated =)

[1] https://gist.githubusercontent.com/anonymous/9fa3dc34b3d59bbd860e8b81ea18a89f/raw/508aaad6abff4ee858b4b584fe6d2405d27dfd7c/badly-scaled.lp

-Ryan
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://list.coin-or.org/pipermail/cbc/attachments/20180209/bca1adb5/attachment.html>