An effective heuristic for large-scale fault-tolerant k-median problem
- 73 Downloads
We address a general fault-tolerant version of the k-median problem on a network. Unlike the original k-median, the objective is to find k nodes (medians or facilities) of a network, assign each non-median node (customer) to \(r_j\) distinct medians, and each median nodes to \(r_j-1\) other medians so as to minimize the overall assignment cost. The problem can be considered in context of the so-called reliable facility location, where facilities once located may be subject to failures. Hedging against possible disruptions, each customer is assigned to multiple distinct facilities. We propose a fast and effective heuristic rested upon consecutive searching for lower and upper bounds for the optimal value. The procedure for finding lower bounds is based on a Lagrangian relaxation and a specialized effective subgradient algorithm for solving the corresponding dual problem. The information on dual variables is then used by a core heuristic in order to determine a set of primal variables to be fixed. The effectiveness and efficiency of our approach are demonstrated in a computational experiment on large-scale problem instances taken from TSPLIB. We show that the proposed algorithm is able to fast find near-optimal solutions to problem instances with almost 625 million decision variables (on networks with up to 24978 vertices).
Keywordsk-Median problem Reliable facility location Core selection Fault-tolerant facility location Lagrangian relaxation Disruptions
This study was funded by Russian Foundation of Basic Research, Project No. 18-07-01037.
Compliance with ethical standards
Conflict of interest
All the authors (I. Vasilyev, A. V. Ushakov, N. Maltugueva and A. Sfroza) declare that they have no conflict of interest.
This article does not contain any studies with human participants or animals performed by any of the authors.
- Bhattacharya S, Chalermsook P, Mehlhorn K, Neumann A (2014) New approximability results for the robust \(k\)-median problem. In: Ravi R, Gørtz IL (eds) Algorithm theory—SWAT 2014: 14th Scandinavian symposium and workshops, Copenhagen, Denmark, July 2–4, 2014. Proceedings. Springer, Cham, pp 50–61. https://doi.org/10.1007/978-3-319-08404-6_5 Google Scholar
- Byrka J, Srinivasan A, Swamy C (2010) Fault-tolerant facility location: a randomized dependent lp-rounding algorithm. In: Eisenbrand F, Shepherd FB (eds) Integer programming and combinatorial optimization: 14th international conference, IPCO 2010, Lausanne, Switzerland, June 9–11, 2010. Proceedings. Springer, Berlin, pp 244–257. https://doi.org/10.1007/978-3-642-13036-6_19 Google Scholar
- Geoffrion AM (1974) Lagrangean relaxation for integer programming. In: Balinski ML (ed) Approaches to integer programming, mathematical programming studies, vol 2. Springer, Berlin, pp 82–114Google Scholar
- Gomes T, Tapolcai J, Esposito C, Hutchison D, Kuipers F, Rak J, de Sousa A, Iossifides A, Travanca R, André J, Jorge L, Martins L, Ugalde PO, Pašić A, Pezaros D, Jouet S, Secci S, Tornatore M (2016) A survey of strategies for communication networks to protect against large-scale natural disasters. In: Proceedings of 8th international workshop on resilient networks design and modeling (RNDM). Gdańsk University of Technology, Gdańsk, pp 11–22. https://doi.org/10.1109/RNDM.2016.7608263
- Rybicki B, Byrka J (2015) Improved approximation algorithm for fault-tolerant facility placement. In: Bampis E, Svensson O (eds) Approximation and online algorithms: 12th international workshop, WAOA 2014, Wrocław, Poland, Sept 11–12, 2014, Revised selected papers. Springer, Cham, pp 59–70. https://doi.org/10.1007/978-3-319-18263-6_6 Google Scholar
- Snyder LV, Daskin MS (2005) Reliability models for facility location: the expected failure cost case. Transp Sci 39(3):400–416Google Scholar
- Vasilyev I, Ushakov A (2017) A shared memory parallel heuristic algorithm for the large-scale \(p\)-median problem. In: Sforza A, Sterle C (eds) Optimization and decision science: methodologies and applications: ODS, Sorrento, Italy, Sept 4–7, 2017. Springer, Cham, pp 295–302. https://doi.org/10.1007/978-3-319-67308-0-30 Google Scholar