Abstract
We investigate a metric facility location problem in a distributed setting. In this problem, we assume that each point is a client as well as a potential location for a facility and that the opening costs for the facilities and the demands of the clients are uniform. The goal is to open a subset of the input points as facilities such that the accumulated cost for the whole point set, consisting of the opening costs for the facilities and the connection costs for the clients, is minimized.
We present a randomized distributed algorithm that computes in expectation an \({\mathcal {O}}(1)\)-approximate solution to the metric facility location problem described above. Our algorithm works in a synchronous message passing model, where each point is an autonomous computational entity that has its own local memory and that communicates with the other entities by message passing. We assume that each entity knows the distance to all the other entities, but does not know any of the other pairwise distances. Our algorithm uses three rounds of all-to-all communication with message sizes bounded to \(\mathcal{O}(\log(n))\) bits, where n is the number of input points.
We extend our distributed algorithm to constant powers of metric spaces. For a metric exponent ℓ≥1, we obtain a randomized \({\mathcal {O}}(1)\)-approximation algorithm that uses three rounds of all-to-all communication with message sizes bounded to \(\mathcal{O}(\log(n))\) bits.
Similar content being viewed by others
Notes
Note that it is even possible that we do not open a facility at x j . This happens if x j has another facility in its neighborhood whose first phase bit set to 1 has a smaller index than x j ’s first phase bit set to 1. An important part of our analysis is to control this “chaining effect”.
Note that it would be sufficient for each point to send only the index of the first phase bit set to 1. However, since each point has to send its ID as well, the message size cannot be reduced to o(logn). Therefore, for simplicity of description, we let each point send all its phase bits to all the other points in the first part of the algorithm.
References
Arya, V., Garg, N., Khandekar, R., Meyerson, A., Munagala, K., Pandit, V.: Local search heuristics for k-median and facility location problems. SIAM J. Comput. 33(3), 544–562 (2004)
Bădoiu, M., Czumaj, A., Indyk, P., Sohler, C.: Facility location in sublinear time. In: Proceedings of the 32nd Annual International Colloquium on Automata, Languages and Programming (ICALP’05), vol. 3580, pp. 866–877. Springer, Berlin (2005)
Byrka, J., Aardal, K.: An optimal bifactor approximation algorithm for the metric uncapacitated facility location problem. SIAM J. Comput. 39(6), 2212–2231 (2010)
Charikar, M., Guha, S.: Improved combinatorial algorithms for facility location problems. SIAM J. Comput. 34(4), 803–824 (2005)
Chudak, F.A., Shmoys, D.B.: Improved approximation algorithms for the uncapacitated facility location problem. SIAM J. Comput. 33(1), 1–25 (2003)
Gfeller, B., Vicari, E.: A randomized distributed algorithm for the maximal independent set problem in growth-bounded graphs. In: Proceedings of the 26th Annual ACM SIGACT-SIGOPS Symposium on Principles of Distributed Computing (PODC’07), pp. 53–60. Association for Computing Machinery, New York (2007)
Guha, S., Khuller, S.: Greedy strikes back: Improved facility location algorithms. J. Algorithms 31(1), 228–248 (1999)
Jain, K., Mahdian, M., Markakis, E., Saberi, A., Vazirani, V.V.: Greedy facility location algorithms analyzed using dual fitting with factor-revealing LP. J. ACM (JACM) 50(6), 795–824 (2003)
Jain, K., Mahdian, M., Saberi, A.: A new greedy approach for facility location problems. In: Proceedings of the 34th Annual ACM Symposium on Theory of Computing (STOC’02), pp. 731–740. Association for Computing Machinery, New York (2002)
Jain, K., Vazirani, V.V.: Approximation algorithms for metric facility location and k-median problems using the primal-dual schema and Lagrangian relaxation. J. ACM (JACM) 48(2), 274–296 (2001)
Korupolu, M.R., Plaxton, C.G., Rajaraman, R.: Analysis of a local search heuristic for facility location problems. J. Algorithms 37(1), 146–188 (2000)
Lynch, N.A.: Distributed Algorithms. Morgan Kaufmann, San Mateo (1996)
Mahdian, M., Ye, Y., Zhang, J.: Approximation algorithms for metric facility location problems. SIAM J. Comput. 36(2), 411–432 (2006)
Mettu, R.R., Plaxton, C.G.: The online median problem. SIAM J. Comput. 32(3), 816–832 (2003)
Moscibroda, T., Wattenhofer, R.: Facility location: Distributed approximation. In: Proceedings of the 24th Annual ACM SIGACT-SIGOPS Symposium on Principles of Distributed Computing (PODC’05), pp. 108–117. Association for Computing Machinery, New York (2005)
Pandit, S., Pemmaraju, S.V.: Return of the primal-dual: Distributed metric facility location. In: Proceedings of the 28th Annual ACM Symposium on Principles of Distributed Computing (PODC’09), pp. 180–189. Association for Computing Machinery, New York (2009)
Peleg, D.: Distributed Computing: A Locality-Sensitive Approach. SIAM Monographs on Discrete Mathematics and Applications (2000)
Shmoys, D.B., Tardos, É., Aardal, K.: Approximation algorithms for facility location problems. In: Proceedings of the 29th Annual ACM Symposium on Theory of Computing (STOC’97), pp. 265–274. Association for Computing Machinery, New York (1997)
Sviridenko, M.: An improved approximation algorithm for the metric uncapacitated facility location problem. In: Proceedings of the 9th International Conference on Integer Programming and Combinatorial Optimization (IPCO’02), pp. 240–257. Springer, Berlin (2002)
Thorup, M.: Quick k-median, k-center, and facility location for sparse graphs. SIAM J. Comput. 34(2), 405–432 (2004)
Author information
Authors and Affiliations
Corresponding author
Additional information
Partially supported by DFG grant Me 872-8/3.
A preliminary version of this paper appeared in the Proceedings of the 18th ACM Symposium on Parallelism in Algorithms and Architectures (SPAA ’06), pages 237–243, 2006.
Appendix: Proof of Lemma 2
Appendix: Proof of Lemma 2
In the proof, we use a modified version of the Mettu-Plaxton algorithm [14]. The original Mettu-Plaxton algorithm computes for each point x i ∈X a special value, which the authors call the radius of x i . Then, it goes through all the points in X in non-decreasing order of their radii and opens a facility at a point x i ∈X if x i has no open facility in the ball whose center is x i and whose radius is twice the radius of x i . Our modified version works exactly as the Mettu-Plaxton algorithm except that, in the first step, it computes, for each point x i ∈X, the radius r i that satisfies (1), instead of the original radius proposed by Mettu and Plaxton [14]. A pseudocode listing of this algorithm is given by Algorithm A.1.
We will first show that this modified Mettu-Plaxton algorithm yields a constant-factor approximation. Based on this result, we will then prove that the sum of the exponentiated radii approximates the optimal cost FacLoc∗(X,f,ℓ) within a constant factor.
Let F MP be the set of open facilities computed by the modified Mettu-Plaxton algorithm. In the following, we will show that \(\operatorname {FacLoc}(X,F_{\textsc{MP}},f,\ell) \le 3^{\ell}\cdot\allowbreak \mathrm {FacLoc}^{*}(X,f,\ell)\). The argumentation is basically the same as in [14]. Only a few minor adaptations to our scenario have been made.
Proposition 1
For any point x i ∈X, there exists an open facility x j ∈F MP such that r j ≤r i and D(x i ,x j )≤2⋅r i .
Proof
If there is no such open facility x j with r j ≤r i in \(\mathcal{B}(x_{i}, 2 \cdot r_{i})\), then we open a facility at x i and x i belongs to F MP . □
Proposition 2
Let x i and x j be distinct open facilities in F MP . Then, we have D(x i ,x j )>2⋅max{r i ,r j }.
Proof
Without loss of generality, we assume that r j ≤r i and x j is previous to x i in the sorted sequence processed by the algorithm. It follows that \(x_{j} \notin \mathcal{B}(x_{i},2 \cdot r_{i})\). Otherwise, the point x i would not be an open facility. Thus, we have
□
For any point x j ∈X and an arbitrary set of open facilities F′⊆X, let
Proposition 3
For an arbitrary set of open facilities F′⊆X, we have
Proof
Due to the definition of \(\operatorname {charge}(\cdot,\cdot)\) and Equation (1), we get
□
Proposition 4
Let x j ∈X be any point, let F′⊆X be an arbitrary set of open facilities, and let x i ∈F′ be any open facility. If we have D(x j ,x i )=D(x j ,F′), then \(\operatorname {charge}(x_{j},F') \ge \max\{ r_{i}^{\ell},\mathrm {D}(x_{j},x_{i})^{\ell}\}\).
Proof
If \(x_{j} \notin \mathcal{B}(x_{i},r_{i})\), then
Otherwise, we have
□
Proposition 5
Let x j ∈X be any point, and let x i be any open facility in F MP . If \(x_{j} \in \mathcal{B}(x_{i},r_{i})\), then \(\operatorname {charge}(x_{j},F_{\textsc{MP}}) \le r_{i}^{\ell}\).
Proof
By Proposition 2, there is no open point x m ∈F MP such that we have i≠m and \(x_{j} \in \mathcal{B}(x_{m},r_{m})\). Since D(x j ,F MP )≤D(x j ,x i ), we obtain
□
Proposition 6
Let x j ∈ be any point, and let x i be any open facility in F MP . If \(x_{j} \notin \mathcal{B}(x_{i},r_{i})\), then we have \(\operatorname {charge}(x_{j},F_{\textsc{MP}}) \le \mathrm {D}(x_{j},x_{i})^{\ell}\).
Proof
The correctness of the assertion follows immediately, unless there is an open facility x m ∈F MP such that \(x_{j} \in \mathcal{B}(x_{m},r_{m})\). If such an open facility x m exists, then Propositions 2 and 5 imply D(x i ,x m )>2⋅max{r i ,r m } and \(\operatorname {charge}(x_{j},F_{\textsc{MP}}) \le r_{m}^{\ell}\). Furthermore, by triangle inequality, we obtain
which proves \(\operatorname {charge}(x_{j},F_{\textsc{MP}}) \le r_{m}^{\ell}\le \mathrm {D}(x_{j},x_{i})^{\ell}\). □
Proposition 7
For any point x j ∈X and an arbitrary set of open facilities F′⊆X, we have \(\operatorname {charge}(x_{j},F_{\textsc{MP}}) \le 3^{\ell}\cdot \operatorname {charge}(x_{j},F')\).
Proof
Let x i be some open facility in F′ such that we have D(x j ,x i )=D(x j ,F′). By Proposition 1, there exists a facility x m ∈F MP such that we have r m ≤r i and D(x i ,x m )≤2⋅r i .
If \(x_{j} \in \mathcal{B}(x_{m},r_{m})\), then we get \(\operatorname {charge}(x_{j},F_{\textsc{MP}}) \le r_{m}^{\ell}\) by Proposition 5. Since Proposition 4 implies \(\operatorname {charge}(x_{j},F') \ge r_{i}^{\ell}\), we can conclude
This proves the assertion in case that we have \(x_{j} \in \mathcal{B}(x_{m},r_{m})\).
If \(x_{j} \notin \mathcal{B}(x_{m},r_{m})\), then \(\operatorname {charge}(x_{j},F_{\textsc{MP}}) \le \mathrm {D}(x_{j},x_{m})^{\ell}\) by Proposition 6. Thus, by triangle inequality, we get
Now, the assertion follows by Proposition 4. □
Lemma 11
\(\operatorname {FacLoc}(X,F_{\textsc{MP}},f,\ell) \le 3^{\ell}\cdot \mathrm {FacLoc}^{*}(X,f,\ell)\)
Proof
The correctness of the assertion follows from Propositions 3 and 7. □
Based on the results above, we can prove Lemma 2. We first prove the lower bound and then the upper bound. The argumentation is basically the same as in [2]. Only a few minor adaptations to our scenario have been made.
Lower bound
Let F MP be the set of open facilities computed by the modified Mettu-Plaxton algorithm. Then, it follows from Proposition 1 that
Next, we show that we also have
Due to Proposition 2, each point x i ∈X is contained in at most one ball \(\mathcal{B}(x_{j},r_{j})\) for some open facility x j ∈F MP . Furthermore, for any point \(x_{m} \in \mathcal{B}(x_{j},r_{j})\), we must have r j ≤2⋅r m . Let us assume that this assumption is false and r m <r j /2. Then, we have that x m is previous to x j in the ordered sequence processed by the algorithm. Furthermore, the algorithm either opens a facility at x m or there is an open facility in the ball \(\mathcal{B}(x_{m},2\cdot r_{m})\). Since
the modified Mettu-Plaxton algorithm would not open a facility at x j , which is a contradiction. Hence, we obtain
which proves inequality (7). Due to inequalities (6) and (7), we get
Upper bound
Due to Lemma 11, we know that
Thus, to prove the upper bound, it remains to show that
Due to Proposition 3, we have
where δ(j) denotes the index of the facility in F MP that is closest to x j . Thus, if we can show that
then we are done. It is sufficient to prove
because this implies \(\max\{ r_{\delta(j)}^{\ell},\mathrm {D}(x_{j},x_{\delta(j)})^{\ell}\} \ge r_{j}^{\ell}/2^{\ell}\) and Inequality (8) follows. We prove the correctness of Inequality (9) by contradiction. Hence, we assume that
We can easily prove by induction that 2ℓ−1⋅(a ℓ+b ℓ)≥(a+b)ℓ for any a,b≥0. Thus, we obtain
which, in turn, would imply \(\mathcal{B}(x_{\delta (j)}, r_{\delta (j)}) \subseteq \mathcal{B}(x_{j}, r_{j})\). Furthermore, by applying triangle inequality and 2ℓ−1⋅(a ℓ+b ℓ)≥(a+b)ℓ for an a,b≥0, we get
as upper bound on the exponentiated distance between x j and any point \(x_{m} \in \mathcal{B}(x_{\delta (j)}, r_{\delta (j)})\). Now, we obtain
which is a contradiction because the definition of r j requires
It follows that inequality (9) is true, which was the only thing left to prove the assertion of the lemma.
Rights and permissions
About this article
Cite this article
Gehweiler, J., Lammersen, C. & Sohler, C. A Distributed O(1)-Approximation Algorithm for the Uniform Facility Location Problem. Algorithmica 68, 643–670 (2014). https://doi.org/10.1007/s00453-012-9690-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00453-012-9690-y