Deep optimization for spectrum repacking

Over 13 months in 2016--17 the U.S. Federal Communications Commission conducted an "incentive auction" to repurpose radio spectrum from broadcast television to wireless internet. In the end, the auction yielded $19.8 bn, $10.05 bn of which was paid to 175 broadcasters for voluntarily relinquishing their licenses across 14 Ultra High Frequency (UHF) channels. Stations that continued broadcasting were assigned potentially new channels to fit as densely as possible into the channels that remained. The government netted more than $7 bn (used to pay down the national debt) after covering costs (including retuning). A crucial element of the auction design was the construction of a solver, dubbed SAT-based Feasibility Checker (SATFC), that determined whether sets of stations could be "repacked" in this way; it needed to run every time a station was given a price quote. This paper describes the process by which we built SATFC. We adopted an approach we dub "deep optimization," taking a data-driven, highly parametric, and computationally intensive approach to solver design. More specifically, to build SATFC we designed software that could pair both complete and local-search SAT-encoded feasibility checking with a wide range of domain-specific techniques, such as constraint graph decomposition and novel caching mechanisms that allow for reuse of partial solutions from related, solved problems. We then used automatic algorithm configuration techniques to construct a portfolio of 8 complementary algorithms to be run in parallel, aiming to achieve good performance on instances that arose in proprietary auction simulations. To evaluate the impact of our solver in this paper, we built an open-source reverse auction simulator. We found that within the short time budget required in practice, SATFC solved more than 95% of the problems it encountered. Furthermore, the incentive auction paired with SATFC produced nearly optimal allocations in a restricted setting and substantially outperformed other alternatives at national scale.

[1]  Carlo Mannino,et al.  Models and solution techniques for frequency assignment problems , 2003, 4OR.

[2]  Martin Gebser,et al.  clasp : A Conflict-Driven Answer Set Solver , 2007, LPNMR.

[3]  Kevin Leyton-Brown,et al.  SATzilla: Portfolio-based Algorithm Selection for SAT , 2008, J. Artif. Intell. Res..

[4]  Yuri Malitsky,et al.  ISAC - Instance-Specific Algorithm Configuration , 2010, ECAI.

[5]  Paul Milgrom,et al.  Putting Auction Theory to Work , 2004 .

[6]  Bart Selman,et al.  Algorithm portfolios , 2001, Artif. Intell..

[7]  Kevin Leyton-Brown,et al.  Hydra: Automatically Configuring Algorithms for Portfolio-Based Selection , 2010, AAAI.

[8]  Yuri Malitsky,et al.  Instance-specific algorithm configuration , 2014, Constraints.

[9]  Holger H. Hoos,et al.  Programming by optimization , 2012, Commun. ACM.

[10]  Ilya Segal,et al.  Deferred-acceptance auctions and radio spectrum reallocation , 2014, EC.

[11]  Pascal Van Hentenryck,et al.  Aeon: Synthesizing Scheduling Algorithms from High-Level Models , 2009 .

[12]  Paul R. Milgrom.  Putting Auction Theory To Work: The Churchill Lectures in Economics , 2004 .

[13]  Carlos Ansótegui,et al.  A Gender-Based Genetic Algorithm for the Automatic Configuration of Algorithms , 2009, CP.

[14]  Leslie Pérez Cáceres,et al.  The irace package: Iterated racing for automatic algorithm configuration , 2016 .

[15]  Martin C. Cooper,et al.  Variable and value elimination in binary constraint satisfaction via forbidden patterns , 2015, J. Comput. Syst. Sci..

[16]  U. Doraszelski,et al.  Ownership Concentration and Strategic Supply Reduction , 2016 .

[17]  Shengwu Li.  Obviously Strategy-Proof Mechanisms , 2017 .

[18]  Samy Bengio,et al.  Understanding deep learning requires rethinking generalization , 2016, ICLR.

[19]  Olivier Roussel,et al.  The International SAT Solver Competitions , 2012, AI Mag..

[20]  Ilya Segal,et al.  Assessing Economic Outcomes in Simulated Reverse Clock Auctions for Radio Spectrum , 2017, ArXiv.

[21]  Ashiqur R. KhudaBukhsh,et al.  SATenstein: automatically building local search SAT solvers from components , 2009, IJCAI 2009.

[22]  Douglas R. Smith,et al.  Synthesis of efficient constraint-satisfaction programs , 2001, The Knowledge Engineering Review.

[23]  Kevin Leyton-Brown,et al.  Sequential Model-Based Optimization for General Algorithm Configuration , 2011, LION.

[24]  Design of the Reverse Auction in the Broadcast Incentive Auction , 2015 .

[25]  Luca Di Gaspero,et al.  EasySyn++: A Tool for Automatic Synthesis of Stochastic Local Search Algorithms , 2007, SLS.

[26]  Michael Kearns,et al.  A Computational Study of Feasible Repackings in the FCC Incentive Auctions , 2014, ArXiv.

[27]  Thomas Stützle,et al.  AClib: A Benchmark Library for Algorithm Configuration , 2014, LION.

[28]  Kevin Leyton-Brown,et al.  Solving the Station Repacking Problem , 2016, AAAI.