Feature-based algorithm selection for constrained continuous optimisation

With this paper, we contribute to the growing research area of feature-based analysis of bio-inspired computing. In this research area, problem instances are classified according to different features of the underlying problem in terms of their difficulty of being solved by a particular algorithm. We investigate the impact of different sets of evolved instances for building prediction models in the area of algorithm selection. Building on the work of Poursoltan and Neumann [1], [2], we consider how evolved instances can be used to predict the best performing algorithm for constrained continuous optimisation from a set of bio-inspired computing methods, namely high performing variants of differential evolution, particle swarm optimization, and evolution strategies. Our experimental results show that instances evolved with a multi-objective approach in combination with random instances of the underlying problem allow to build a model that accurately predicts the best performing algorithm for a wide range of problem instances.

[1]  Carlos A. Coello Coello,et al.  Constraint-handling in nature-inspired numerical optimization: Past, present and future , 2011, Swarm Evol. Comput..

[2]  Yong Wang,et al.  A hybrid multi-swarm particle swarm optimization to solve constrained optimization problems , 2009, Frontiers of Computer Science in China.

[3]  Kevin Leyton-Brown,et al.  Performance Prediction and Automated Tuning of Randomized and Parametric Algorithms , 2006, CP.

[4]  Tetsuyuki Takahama,et al.  Constrained optimization by the ε constrained differential evolution with an archive and gradient-based mutation , 2010, IEEE Congress on Evolutionary Computation.

[5]  Kate Smith-Miles,et al.  Towards insightful algorithm selection for optimisation using meta-learning concepts , 2008, 2008 IEEE International Joint Conference on Neural Networks (IEEE World Congress on Computational Intelligence).

[6]  Dirk V. Arnold,et al.  A (1+1)-CMA-ES for constrained optimisation , 2012, GECCO '12.

[7]  Frank Neumann,et al.  A Feature-Based Analysis on the Impact of Set of Constraints for e-Constrained Differential Evolution , 2015, ArXiv.

[8]  Bernd Bischl,et al.  Algorithm selection based on exploratory landscape analysis and cost-sensitive learning , 2012, GECCO '12.

[9]  P. Suganthan,et al.  Problem Definitions and Evaluation Criteria for the CEC 2010 Competition on Constrained Real- Parameter Optimization , 2010 .

[10]  R. Geoff Dromey,et al.  An algorithm for the selection problem , 1986, Softw. Pract. Exp..

[11]  Frank Neumann,et al.  A Feature-Based Comparison of Evolutionary Computing Techniques for Constrained Continuous Optimisation , 2015, ICONIP.

[12]  Bogdan Filipic,et al.  DEMO: Differential Evolution for Multiobjective Optimization , 2005, EMO.

[13]  Frank Neumann,et al.  A Feature-based analysis on the impact of linear constraints for ε-constrained differential evolution , 2014, 2014 IEEE Congress on Evolutionary Computation (CEC).

[14]  Mario A. Muñoz,et al.  A Meta-learning Prediction Model of Algorithm Performance for Continuous Optimization Problems , 2012, PPSN.

[15]  Ross J. W. James,et al.  A Knowledge Discovery Approach to Understanding Relationships between Scheduling Problem Structure and Heuristic Performance , 2009, LION.

[16]  Yoav Shoham,et al.  Empirical hardness models: Methodology and a case study on combinatorial auctions , 2009, JACM.

[17]  D. Marquardt.  An Algorithm for Least-Squares Estimation of Nonlinear Parameters , 1963 .