I know that A * and CSP are different paradigms. I made several projects with both. Both are also used in the planning / scheduling problem. A * is easy to implement, simpler, and the variants are interesting when integrated in real-time systems. The CSP (with BAB engine) seems more suitable, but may require more time to be formalized, mapped to an existing solver. So, maybe both solutions should be tested and evaluated (development, complexity, maintainability, evolutivity,...)
Thanks, Philippe > Le 13 mars 2018 à 16:10, Christian Schulte <cschu...@kth.se> a écrit : > > A* and how BAB works in Gecode is very different. A* is typically based on > representing the problem by states and then coming up with the right > operators and functions, etc. > > Gecode uses constraint programming which has a very different way of modeling > problems. That's why I would suggest to really try a little bit of how models > with constraint programming look like. If you want to get an idea on a more > high-level you could try a modeling language such as MiniZinc: > http://www.minizinc.org/ > > Cheers > Christian > > -- > Christian Schulte, https://chschulte.github.io/ > Professor of Computer Science, KTH, cschu...@kth.se > Expert Researcher, RISE SICS, christian.schu...@ri.se > > > -----Original Message----- > From: aqwzsxaqw...@orange.fr [mailto:aqwzsxaqw...@orange.fr] > Sent: Tuesday, March 13, 2018 16:06 > To: users@gecode.org; Christian Schulte <cschu...@kth.se> > Subject: RE: RE: [gecode-users] BAB implementation > > > Hi, > > Definitively not a homework. I have to make a technical choice to solve a > planification problem with task/resources/scheduling. > A * will be homogeneous with the rest of the software but CSP with a BAB > engine seems clearly more appropriate to me. > It has been proved in 1984 by Nau/Kumar/Kanal that A* is a special case of a > general BAB formulation and I don't see any reason why it would not be the > case with Gecode, but i was asking some advice to decide ! > Moreover, i will maybe have some questions about coding rules in gecode > relative to embedded system. > > Philippe > > > ---------------- > Le 13/03/2018, à 15:16, "Christian Schulte" <cschu...@kth.se> a écrit : > > Hi, I think you will have to do some reading. Smells like homework to me. > Christian > > -- > Christian Schulte, https://chschulte.github.io/ Professor of Computer > Science, KTH, cschu...@kth.se Expert Researcher, RISE SICS, > christian.schu...@ri.se > > > -----Original Message----- > From: users-boun...@gecode.org [mailto:users-boun...@gecode.org] On Behalf Of > aqwzsxaqw...@orange.fr > Sent: Tuesday, March 13, 2018 15:02 > To: users@gecode.org > Subject: [gecode-users] BAB implementation > > Hi all, > > This is maybe a general question about the B&B and the A* algorithm, But I > was wondering what was the advantage, force, weakness of the Gecode BAB > engine over a A*. > > Thanks > > _______________________________________________ > Gecode users mailing list > users@gecode.org > https://www.gecode.org/mailman/listinfo/gecode-users _______________________________________________ Gecode users mailing list users@gecode.org https://www.gecode.org/mailman/listinfo/gecode-users
