Why Asynchronous Parallel Evolution is the Future of Hyper-heuristics: A CDCL SAT Solver Case Study

Evolutionary Algorithms (EAs) are inherently parallel due to their ability to simultaneously evaluate the fitness of individuals. Synchronous Parallel EAs (SPEAs) leverage this with the intent to gain significant speed-ups when executed on multiple processors. However, many important problem classes lead to large variations in fitness evaluation times, such as is often the case in hyper-heuristics where the time complexity of executing one individual may differ greatly from that of another. Asynchronous Parallel EAs (APEAs) omit the generational synchronization step of traditional EAs which work in well-defined cycles. They can provide scalability improvements proportional to the variation in fitness evaluation times of the evolved individuals, and therefore should be considered for use in hyper-heuristics. This paper provides an empirical analysis of the improvements obtained by applying APEAs, compared to SPEAs, on a case study involving the evolution of conflict-driven clause learning Boolean satisfiability solvers, demonstrating that APEAs are the future of hyper-heuristics.

[1]  A. E. Eiben,et al.  Introduction to Evolutionary Computing , 2003, Natural Computing Series.

[2]  Mark Harman,et al.  Applying Genetic Improvement to MiniSAT , 2013, SSBSE.

[3]  Armando Tacchella,et al.  Theory and Applications of Satisfiability Testing: 6th International Conference, Sat 2003, Santa Margherita Ligure, Italy, May 5-8 2003: Selected Revised Papers (Lecture Notes in Computer Science, 2919) , 2004 .

[4]  Alex S. Fukunaga,et al.  Automated Discovery of Local Search Heuristics for Satisfiability Testing , 2008, Evolutionary Computation.

[5]  Lothar Thiele,et al.  Multiobjective genetic programming: reducing bloat using SPEA2 , 2001, Proceedings of the 2001 Congress on Evolutionary Computation (IEEE Cat. No.01TH8546).

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

[7]  Enrique Alba,et al.  Analyzing synchronous and asynchronous parallel distributed genetic algorithms , 2001, Future Gener. Comput. Syst..

[8]  Enrique Alba,et al.  Parallelism and evolutionary algorithms , 2002, IEEE Trans. Evol. Comput..

[9]  Alex S. Fukunaga,et al.  Evolving Local Search Heuristics for SAT Using Genetic Programming , 2004, GECCO.

[10]  Riccardo Poli,et al.  A Field Guide to Genetic Programming , 2008 .

[11]  Alex S. Fukunaga,et al.  Massively parallel evolution of SAT heuristics , 2009, 2009 IEEE Congress on Evolutionary Computation.

[12]  Andrew Philippides,et al.  Tool sequence optimization using synchronous and asynchronous parallel multi-objective evolutionary algorithms with heterogeneous evaluations , 2013, 2013 IEEE Congress on Evolutionary Computation.

[13]  Daniel R. Tauritz,et al.  Asynchronous Parallel Evolutionary Algorithms: Leveraging Heterogeneous Fitness Evaluation Times for Scalability and Elitist Parsimony Pressure , 2015, GECCO.

[14]  Bastien Chopard,et al.  Parallel Genetic Programming and its Application to Trading Model Induction , 1997, Parallel Comput..

[15]  Marc Schoenauer,et al.  Asynchronous master/slave moeas and heterogeneous evaluation costs , 2012, GECCO '12.

[16]  Daniel R. Tauritz,et al.  Hyper-Heuristics: A Study On Increasing Primitive-Space , 2015, GECCO.

[17]  Armin Biere,et al.  Evaluating CDCL Restart Schemes , 2018, POS@SAT.

[18]  Enrique Alba,et al.  Parallel evolutionary algorithms can achieve super-linear performance , 2002, Inf. Process. Lett..

[19]  Marius Thomas Lindauer,et al.  SpySMAC: Automated Configuration and Performance Analysis of SAT Solvers , 2015, SAT.

[20]  Riccardo Poli,et al.  Generating SAT Local-Search Heuristics Using a GP Hyper-Heuristic Framework , 2007, Artificial Evolution.

[21]  Marc Schoenauer,et al.  Asynchronous Evolutionary Multi-Objective Algorithms with heterogeneous evaluation costs , 2011, 2011 IEEE Congress of Evolutionary Computation (CEC).

[22]  Armin Biere,et al.  Evaluating CDCL Variable Scoring Schemes , 2015, SAT.

[23]  Kenneth A. De Jong,et al.  Evaluation-Time Bias in Asynchronous Evolutionary Algorithms , 2015, GECCO.

[24]  Niklas Sörensson,et al.  An Extensible SAT-solver , 2003, SAT.

[25]  Daniel R. Tauritz,et al.  A problem configuration study of the robustness of a black-box search algorithm hyper-heuristic , 2014, GECCO.

[26]  Mark Harman,et al.  Using Genetic Improvement and Code Transplants to Specialise a C++ Program to a Problem Class , 2014, EuroGP.

[27]  Julian F. Miller,et al.  What bloat? Cartesian Genetic Programming on Boolean problems , 2003 .

[28]  Enrique Alba,et al.  A study of master-slave approaches to parallelize NSGA-II , 2008, 2008 IEEE International Symposium on Parallel and Distributed Processing.

[29]  Daniel R. Tauritz,et al.  Multi-sample evolution of robust black-box search algorithms , 2014, GECCO.

[30]  You Li,et al.  Optimizing the Initialization of Dynamic Decision Heuristics in DPLL SAT Solvers Using Genetic Programming , 2006, EuroGP.