Currently the best approximation algorithm for the MULTIWAY CUT problem is obtained via the linear program based on geometrical embedding by CKR [1]. Let $U_i$ be those vertices in $V-T$ which is assigned $e_i = \{0, \dots, 1, \dots, 0\}$, that is, $1$ at the $i$-th coordinate and $0$ at all others. My question is: is there always an optimal solution that keeps vertices $U_i$ with $t_i$ (the $i$-th terminal)?
Remark. For the multicommodity-flow-based linear program for the MULTIWAY CUT problem (see [2]), it is not hard to show the existence of an optimal solution. It should be verified with the similar idea as [3], which proves the existence of such an optimal solution on the NODE MULTIWAY CUT problem.
- Calinescu, G., H. Karloff, Y. Rabani. 2000. An improved approximation algorithm for Multiway Cut. J. Comput. System Sci. 60(3) 564–574.
- E. Dahlhaus, D. S. Johnson, C. H. Papadimitriou, P. D. Seymour, and M. Yannakakis, The complexity of multiterminal cuts, SIAM J. Comput. 23 (1994), 864894. [Preliminary version appeared in STOC '92]
- Sylvain Guillemot, FPT algorithms for path-transversal and cycle-transversal problems, Discrete Optimization 8 (2011) 61–71
P.S. I'm aware of Karger's result, but that is irrelevant.