Hi,
I am wondering is it possible to provide a warm start to DAQP BnB?
In my application, I have almost always an integer feasible guess, but I noticed that the timing does not change if I actually pass it to DAQP.
use DAQP via CasADi's conic, which is possible since the latest 3.8.0. release (and not directly via its python interface).
I do not know is it an issue on the CasADi interface side or does DAQP not accept the binary to prune the BnB tree?
Regarding the former, if I pass a warm start to gurobi via CasADi's conic, it gets accepted and I do see an significant improvement in computation times.
Thanks!
Hi,
I am wondering is it possible to provide a warm start to DAQP BnB?
In my application, I have almost always an integer feasible guess, but I noticed that the timing does not change if I actually pass it to DAQP.
use DAQP via CasADi's conic, which is possible since the latest 3.8.0. release (and not directly via its python interface).
I do not know is it an issue on the CasADi interface side or does DAQP not accept the binary to prune the BnB tree?
Regarding the former, if I pass a warm start to gurobi via CasADi's conic, it gets accepted and I do see an significant improvement in computation times.
Thanks!