The focus of the seminar are methods and algorithms for combinatorial optimization. This includes, but not limited to
- convex relaxations and respective convex large-scale optimization methods
- combinatoril heuristics
- classical techniques for polynomially solvable combinatorial problems, such as the min-cost flow problem and its special cases: Linear assignment, max-flow and transportation ones.
Upcoming talk:
9.03.2020, 13:00, Mathematikon B, Room B128
Bogdan Savchynskyy. Weighted vertex-packing problem: LP-relaxation and persistency
Schedule:
16.03.2020, 13:00, Mathematikon B, Room B128
Stefan Haller. Network Simplex Method and reparameterization of the linear assigment problem
Past talks:
16.12.2019, 13:00, Mathematikon B, Room B128
Jakob Schnell. Overview of Primal Heuristics for Quadratic Assignment Problem (Graph Matching)
18.11.2019, 13:00, Mathematikon B, Room B128
Lisa Kruse. Hungarian Method as Primal-Dual Algorithm. Insights into reparametrization of linear assignemt problems
4.11.2019, 13:00, Mathematikon B, Room B128
Siddharth Tourani. Auction Algorithm as a dual coordinate-ascent method
9.09.2019, 13:00, Mathematikon B, Room B128
Bogdan Savchynskyy. Sinkhorn Algorithm and alternating projection method of Bregmann for linear assignment.
Relation to dual block-coordinate-ascent methods for discrete graphical models.