A dynamic program which solves CSP instances in pseudo-polynomial time has been defined in literature. In order to be able to reshuffle sequences for a real-world car manufacturing company using mix banks, we propose a dynamic programming heuristic which considers input sequences of unbounded length and constraints of different types and priorities. We denote this new problem by eCRSP, an extended version of CRSP. We prune states using a dominance rule and final penalty estimations of partial solutions. We verify the performance of the program using the planned sequences that enter the factory, scheduled by a commercially available sequencing tool, sequences that have been manufactured in the plant in the past and instances from CSPLib, a publicly accessible set of constraint satisfaction problems.
--------
Join Zoom Meeting
Meeting ID: 527 278 8807
Note: for people outside D1 interested in listening to this talk, please contact Sándor Kisfaludi-Bak at skisfalu@mpi-inf.mpg.de for the password.