ABSTRACT
The choice of parameter values greatly affects the performance of evolutionary algorithms. Many current parameter tuning approaches require multiple runs of the tuned algorithm, which makes it hard to use them in budget-constrained environments. Recently, an approach to parameter tuning on binary string problems was proposed, which uses machine learning and fitness landscape analysis to transfer knowledge on good parameter choices between similar problem instances. This approach allows using the performance data obtained on simple benchmark problems with different fitness landscape features to quickly choose suitable parameters for a given problem instance.
In this paper, we aim to extend this approach to permutation-based problems by tuning the recently proposed version of the (1 + (λ, λ)) genetic algorithm for permutations-based problems. To do this, we develop a set of fitness landscape features that can be computed for permutations. We collect the algorithm's performance dataset on multiple instances of the W-model benchmark problem layered over the Ham problem for permutations. Finally, we present the preliminary experimental evaluation of the (1 + (λ, λ)) genetic algorithm tuned by the proposed approach.
- Anton Bassin and Maxim Buzdalov. 2020. The (1 + (λ, λ)) Genetic Algorithm for Permutations. In Proceedings of Genetic and Evolutionary Computation Conference Companion. ACM, 1669--1677.Google ScholarDigital Library
- Vincent A Cicirello. 2019. Classification of permutation distance metrics for fitness landscape analysis. In International Conference on Bio-inspired Information and Communication. Springer, 81--97.Google ScholarCross Ref
- Benjamin Doerr and Carola Doerr. 2015. Optimal parameter choices through self-adjustment: Applying the 1/5-th rule in discrete settings. In Proceedings of Genetic and Evolutionary Computation Conference. 1335--1342.Google ScholarDigital Library
- Benjamin Doerr, Carola Doerr, and Franziska Ebel. 2015. From black-box complexity to designing new genetic algorithms. Theoretical Computer Science 567 (2015), 87--104.Google ScholarDigital Library
- Ágoston Endre Eiben, Zbigniew Michalewicz, Marc Schoenauer, and J. E. Smith. 2007. Parameter control in evolutionary algorithms. In Parameter Setting in Evolutionary Algorithms. Number 54 in Studies in Computational Intelligence. Springer, 19--46.Google Scholar
- Agoston E Eiben and Selmar K Smit. 2011. Parameter tuning for configuring and analyzing evolutionary algorithms. Swarm and Evolutionary Computation 1, 1 (2011), 19--31.Google ScholarCross Ref
- Leticia Hernando, Alexander Mendiburu, and Jose A Lozano. 2015. A tunable generator of instances of permutation-based combinatorial optimization problems. IEEE Transactions on Evolutionary Computation 20, 2 (2015), 165--179.Google ScholarDigital Library
- Frank Hutter, Holger H. Hoos, and Kevin Leyton-Brown. 2011. Sequential model-based optimization for general algorithm configuration. In Proceedings of Learning and Intelligent Optimization. Springer, 507--523.Google ScholarDigital Library
- Anja Janković and Carola Doerr. 2019. Adaptive landscape analysis. In Proceedings of the Genetic and Evolutionary Computation Conference Companion. 2032--2035.Google ScholarDigital Library
- Giorgos Karafotias, Mark. Hoogendoorn, and Ágoston E. Eiben. 2015. Parameter Control in Evolutionary Algorithms: Trends and Challenges. IEEE Transactions on Evolutionary Computation 19, 2 (2015), 167--187.Google ScholarDigital Library
- Pascal Kerschke and Heike Trautmann. 2016. The R-package FLACCO for exploratory landscape analysis with applications to multi-objective optimization problems. In 2016 IEEE Congress on Evolutionary Computation. IEEE, 5262--5269.Google ScholarDigital Library
- Pascal Kerschke and Heike Trautmann. 2019. Automated algorithm selection on continuous black-box problems by combining exploratory landscape analysis and machine learning. Evolutionary computation 27, 1 (2019), 99--127.Google Scholar
- Fernando G. Lobo, Cláudio F. Lima, and Zbigniew Michalewicz (Eds.). 2007. Parameter Setting in Evolutionary Algorithms. Number 54 in Studies in Computational Intelligence. Springer.Google Scholar
- Manuel López-Ibáñez, Jérémie Dubois-Lacoste, Leslie Pérez Cáceres, Thomas Stützle, and Mauro Birattari. 2016. The irace package: Iterated racing for automatic algorithm configuration. Operations Research Perspectives 3 (2016), 43--58.Google ScholarCross Ref
- Marie-Eléonore Marmion and Olivier Regnier-Coudert. 2015. Fitness landscape of the factoradic representation on the permutation flowshop scheduling problem. In International Conference on Learning and Intelligent Optimization. Springer, 151--164.Google ScholarCross Ref
- Olaf Mersmann, Bernd Bischl, Heike Trautmann, Mike Preuss, Claus Weihs, and Günter Rudolph. 2011. Exploratory landscape analysis. In Proceedings of the 13th annual conference on Genetic and evolutionary computation. 829--836.Google ScholarDigital Library
- Olaf Mersmann, Mike Preuss, and Heike Trautmann. 2010. Benchmarking evolutionary algorithms: Towards exploratory landscape analysis. In International Conference on Parallel Problem Solving from Nature. Springer, 73--82.Google ScholarCross Ref
- Gabriela Ochoa and Katherine Malan. 2019. Recent Advances in Fitness Landscape Analysis. In Proceedings of the Genetic and Evolutionary Computation Conference Companion (Prague, Czech Republic) (GECCO '19). Association for Computing Machinery, New York, NY, USA, 1077--1094. Google ScholarDigital Library
- Maxim Pikalov and Vladimir Mironovich. 2021. Automated Parameter Choice with Exploratory Landscape Analysis and Machine Learning. In Proceedings of the Genetic and Evolutionary Computation Conference Companion (Lille, France) (GECCO '21). 1982--1985. Google ScholarDigital Library
- Jens Scharnow, Karsten Tinnefeld, and Ingo Wegener. 2004. The analysis of evolutionary algorithms on sorting and shortest path problems. Journal of Mathematical Modelling and Algorithms 3 (2004), 349--366.Google ScholarCross Ref
- Tommaso Schiavinotto and Thomas Stützle. 2007. A review of metrics on permutations for search landscape analysis. Computers & operations research 34, 10 (2007), 3143--3153.Google Scholar
- Jorge Tavares, Francisco B Pereira, and Ernesto Costa. 2008. Multidimensional knapsack problem: A fitness landscape analysis. IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics) 38, 3 (2008), 604--616.Google ScholarDigital Library
- Thomas Weise and Zijun Wu. 2018. Difficult Features of Combinatorial Optimization Problems and the Tunable W-Model Benchmark Problem for Simulating them. In Proceedings of Genetic and Evolutionary Computation Conference Companion. 1769--1776.Google ScholarDigital Library
Index Terms
- Towards landscape-aware parameter tuning for the (1 + (λ, λ)) genetic algorithm for permutations
Recommendations
Parameter Tuning for the Genetic Algorithm Using Landscape Analysis and Machine Learning
Applications of Evolutionary ComputationAbstractThe choice of parameter values in evolutionary algorithms greatly affects their performance. Many popular parameter tuning techniques are limited by the tuning budget for finding a good set of parameter values. Recently, we proposed an approach to ...
Optimal Static and Self-Adjusting Parameter Choices for the $$(1+(\lambda ,\lambda ))$$(1+(ź,ź)) Genetic Algorithm
The $$(1+(\lambda ,\lambda ))$$(1+(ź,ź)) genetic algorithm proposed in Doerr et al. (Theor Comput Sci 567:87---104, 2015) is one of the few examples for which a super-constant speed-up of the expected optimization time through the use of crossover could ...
Theoretical and Empirical Analysis of Parameter Control Mechanisms in the (1 + (λ, λ)) Genetic Algorithm
The self-adjusting (1 + (λ, λ)) GA is the best known genetic algorithm for problems with a good fitness-distance correlation as in OneMax. It uses a parameter control mechanism for the parameter λ that governs the mutation strength and the number of ...
Comments