Good day to all, (repeated because chose wrong email address). I wrote an initial message (through the issues site) asking about the possibility of using the It was pointed out that the: 1. The algorithm is deterministic and that 2. It’s not possible to adapt it to iterate in a unbiased random way. Now I have a couple more questions, namely searching for advice on how to achieve what I want. Context: I’ve modeled certain parts of combinatorial optimization problem as a set of graphs that I’m exploiting to efficiently generate variables on the fly during the optimization routine. A variable in my problem is represented by a path between two nodes in one of those graphs. Specifically, what I want:
For the sampling in the third step I’m currently doing the following (assuming a sample of size X and a population of paths of Y):
Many times, the sample X is really small compared to the population Y). X can be 500-2000, and Y can be 100.000+. My problem is: Since I could potentially do many iterations (and samplings), I would like to have an unbiased sampling method for the Y paths without having to enumerate them all. One option is to do the sampling during the construction of the paths. I have in mind to just use the Another option, although a lot more far-fetched, is to create my own Are there better options to doing this? Thanks, Franco Peschiera _______________________________________________ graph-tool mailing list [hidden email] https://lists.skewed.de/mailman/listinfo/graph-tool |
Free forum by Nabble | Edit this page |