Approximating Minimum Bounded Degree Spanning Trees to within One of Optimal
read more
Citations
Computers And Intractability A Guide To The Theory Of Np Completeness
Graph spanners: A tutorial review
Polynomial Bounds for the Grid-Minor Theorem
Solving Optimization Problems with Diseconomies of Scale via Decoupling
The Kadison-Singer Problem for Strongly Rayleigh Measures and Applications to Asymmetric TSP
References
Computers and Intractability: A Guide to the Theory of NP-Completeness
Approximation Algorithms
Combinatorial optimization. Polyhedra and efficiency.
The ellipsoid method and its consequences in combinatorial optimization
Related Papers (5)
Frequently Asked Questions (17)
Q2. Why do the authors remove degree constraints on vertices with at least two strictly fractional edges?
Due to presence of upper and lower degree constraints, the authors only remove degree constraints on vertices which have at most two strictly fractional edges incident at them.
Q3. What is the standard approach to design approximation algorithms?
A standard approach to design approximation algorithms is to first formulate the problem as an integer program, and then use the linear relaxation of this program as a way to lower-bound the cost of an optimal solution to a minimization problem.
Q4. What is the proof of integrality of linear programming formulations?
The iterative rounding framework can be used to obtain new proofs of integrality of linear programming formulations for many combinatorial optimization problems including matchings in bipartite and general graphs, matroid bases and matroid intersection, submodular flows, etc.
Q5. What is the general strategy for a spanning tree of cost?
Using a polyhedral approach, a general strategy is to construct a spanning tree of cost no more than the optimal value of this linear program, and in which the degree of each vertex v ∈ V is at most Bv +1.
Q6. What is the key insight of the algorithm?
The key insight is that if the algorithm cannot find integral edges to pick, then it can remove/relax one of the degree constraints.
Q7. What is the smallest set in the laminar family L?
S receives x∗e token for each edge e such that S is the smallest set containing both endpoints of e. Let R1, . . . , Rk be the children of S in the laminar family L where k ≥ 0.
Q8. What is the definition of the MBDST problem?
When all degree bounds are two (i.e., Bv = 2 for all v), the MBDST problem specializes to the Minimum Cost Hamiltonian Path problem, and thus even the problem of checking whether there exists a feasible solution is NP-complete.
Q9. How does the algorithm generalize to the setting?
The algorithm also generalizes when there are different degree bounds on different vertices, where the algorithm returns a spanning tree violating the degree bounds by at most two.
Q10. What is the main addition to the proof of Lemma 3.1?
It is a simple iterative algorithm and an extension of Iterative MST Algorithm I. The main addition is Step 2(c) that iteratively relaxes degree constraints.
Q11. What is the simplest way to denote the set of edges in the graph?
Given an undirected graph G = (V, E) and a subset S of vertices, the authors use E(S) = {e ∈ E : |e ∩ S| = 2} to denote the set of edges which have both endpoints in S ⊆ V .
Q12. What is the key feature of the iterative rounding method?
A key feature of the iterative rounding method is to repeat this procedure: solve again the linear program for the residual problem to obtain an optimal extreme point solution (instead of using x∗), and add the edges with the highest fractional value in this new fractional solution to the integral solution.
Q13. What is the optimal extreme point solution to LP-MST(G′)?
If x∗ is an optimal extreme point solution to LP-MST(G), then the residual LP solution x∗res, x restricted to G′ = G\\v remains an optimal extreme point solution to LP-MST(G′).
Q14. What is the degree bound of the spanning tree of cost?
Then T is a spanning tree of cost at most c(x∗) where x∗ is the optimal LP solution to LPMBDST(G, V ), and the degree of vertex v in T is at most Bv + 2 for each vertex v ∈ V .PROOF.
Q15. What is the case when S contains two active vertex?
If S contains two active vertices, the authors have the two extra tokens to assign to S. Recall that S itself does not need any tokens since it has been assigned two tokens from edges in F.
Q16. What is the simplest way to denote the set of edges in F?
For any subset of edges F ⊆ E and vertices S ⊆ V , the authors denote δF(S) to be the set of edges in F which have exactly one endpoint in S.
Q17. What is the recursive solution to the residual problem?
Observe that the residual problem is to find a minimum spanning tree on G′ = G\\v, and the same procedure is applied to solve the residual problem recursively.