Automated Configuration of Genetic Algorithms by Tuning for Anytime Performance

Finding the best configuration of algorithms’ hyperparameters for a given optimization problem is an important task in evolutionary computation. We compare in this work the results of four different hyperparameter tuning approaches for a family of genetic algorithms on 25 diverse pseudo-Boolean optimization problems. More precisely, we compare previously obtained results from a grid search with those obtained from three automated configuration techniques: iterated racing, mixedinteger parallel efficient global optimization, and mixed-integer evolutionary strategies. Using two different cost metrics, expected running time and the area under the empirical cumulative distribution function curve, we find that in several cases the best configurations with respect to expected running time are obtained when using the area under the empirical cumulative distribution function curve as the cost metric during the configuration process. Our results suggest that even when interested in expected running time performance, it might be preferable to use anytime performance measures for the configuration task. We also observe that tuning for expected running time is much more sensitive with respect to the budget that is allocated to the target algorithms.

[1]  Hao Wang,et al.  A new acquisition function for Bayesian optimization based on the moment-generating function , 2017, 2017 IEEE International Conference on Systems, Man, and Cybernetics (SMC).

[2]  K. Dejong,et al.  An analysis of the behavior of a class of genetic adaptive systems , 1975 .

[3]  Thomas Bäck,et al.  A Survey of Evolution Strategies , 1991, ICGA.

[4]  John J. Grefenstette,et al.  Optimization of Control Parameters for Genetic Algorithms , 1986, IEEE Transactions on Systems, Man, and Cybernetics.

[5]  Dogan Corus,et al.  Standard Steady State Genetic Algorithms Can Hillclimb Faster Than Mutation-Only Evolutionary Algorithms , 2017, IEEE Transactions on Evolutionary Computation.

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

[7]  William M. Spears,et al.  Crossover or Mutation? , 1992, FOGA.

[8]  Ofer M. Shir,et al.  Benchmarking discrete optimization heuristics with IOHprofiler , 2019, GECCO.

[9]  Thomas Bäck,et al.  Automatic Configuration of Deep Neural Networks with Parallel Efficient Global Optimization , 2019, 2019 International Joint Conference on Neural Networks (IJCNN).

[10]  Carola Doerr,et al.  A Simple Proof for the Usefulness of Crossover in Black-Box Optimization , 2018, PPSN.

[11]  Thomas Bäck,et al.  Mixed Integer Evolution Strategies for Parameter Optimization , 2013, Evolutionary Computation.

[12]  Thomas Bäck,et al.  Parallel Optimization of Evolutionary Algorithms , 1994, PPSN.

[13]  Carola Doerr,et al.  Towards large scale automated algorithm design by integrating modular benchmarking frameworks , 2021, GECCO Companion.

[14]  Dirk Sudholt,et al.  How Crossover Speeds up Building Block Assembly in Genetic Algorithms , 2014, Evolutionary Computation.

[15]  Tadahiko MURATA,et al.  Positive and negative combination effects of crossover and mutation operators in sequencing problems , 1996, Proceedings of IEEE International Conference on Evolutionary Computation.

[16]  Thomas Stützle,et al.  Stochastic Local Search: Foundations & Applications , 2004 .

[17]  Ruhul A. Sarker,et al.  Multi-operator based evolutionary algorithms for solving constrained optimization problems , 2011, Comput. Oper. Res..

[18]  Thomas Stützle,et al.  F-Race and Iterated F-Race: An Overview , 2010, Experimental Methods for the Analysis of Optimization Algorithms.

[19]  David E. Goldberg,et al.  Genetic Algorithms in Search Optimization and Machine Learning , 1988 .

[20]  Thomas Bäck,et al.  Evolutionary algorithms in theory and practice - evolution strategies, evolutionary programming, genetic algorithms , 1996 .

[21]  Peter I. Frazier,et al.  A Tutorial on Bayesian Optimization , 2018, ArXiv.

[22]  Dirk Thierens,et al.  Optimal mixing evolutionary algorithms , 2011, GECCO '11.

[23]  Hao Wang,et al.  IOHprofiler: A Benchmarking and Profiling Tool for Iterative Optimization Heuristics , 2018, ArXiv.

[24]  Byung Ro Moon,et al.  An empirical study on the synergy of multiple crossover operators , 2002, IEEE Trans. Evol. Comput..

[25]  Dirk Sudholt,et al.  The Benefits of Population Diversity in Evolutionary Algorithms: A Survey of Rigorous Runtime Analyses , 2018, Theory of Evolutionary Computation.

[26]  Heike Trautmann,et al.  Anytime Behavior of Inexact TSP Solvers and Perspectives for Automated Algorithm Selection , 2020, 2020 IEEE Congress on Evolutionary Computation (CEC).

[27]  John H. Holland,et al.  When will a Genetic Algorithm Outperform Hill Climbing , 1993, NIPS.

[28]  Thomas Stützle,et al.  Automatically improving the anytime behaviour of optimisation algorithms , 2014, Eur. J. Oper. Res..

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

[30]  Thomas Bartz-Beielstein,et al.  Sequential parameter optimization , 2005, 2005 IEEE Congress on Evolutionary Computation.

[31]  Nando de Freitas,et al.  Taking the Human Out of the Loop: A Review of Bayesian Optimization , 2016, Proceedings of the IEEE.

[32]  Carola Doerr,et al.  Hyper-parameter tuning for the (1 + (λ, λ)) GA , 2019, GECCO.

[33]  Ingo Rechenberg,et al.  Evolution Strategy: Nature’s Way of Optimization , 1989 .

[34]  Thomas Bäck,et al.  IOHanalyzer: Performance Analysis for Iterative Optimization Heuristic , 2020, ArXiv.

[35]  Per Kristian Lehre,et al.  Black-Box Search by Unbiased Variation , 2010, GECCO '10.

[36]  Thomas Weise,et al.  Selecting a diverse set of benchmark instances from a tunable model problem for black-box discrete optimization algorithms , 2020, Appl. Soft Comput..

[37]  Hao Wang,et al.  Cooling Strategies for the Moment-Generating Function in Bayesian Global Optimization , 2018, 2018 IEEE Congress on Evolutionary Computation (CEC).