A meta-learning prediction model of algorithm performance for continuous optimization problems

Authors:
Mario A. Muñoz;Michael Kirley;Saman K. Halgamuge
Affiliations:
Department of Mechanical Engineering, The University of Melbourne, Parkville, Victoria, Australia;Department of Computing and Information Systems, The University of Melbourne, Parkville, Victoria, Australia;Department of Mechanical Engineering, The University of Melbourne, Parkville, Victoria, Australia
Venue:
PPSN'12 Proceedings of the 12th international conference on Parallel Problem Solving from Nature - Volume Part I
Year:
2012

Citing 17
Cited 2

Fitness landscapes and evolvability

Evolutionary Computation
Fitness Distance Correlation as a Measure of Problem Difficulty for Genetic Algorithms

Proceedings of the 6th International Conference on Genetic Algorithms
The dispersion metric and the CMA evolution strategy

Proceedings of the 8th annual conference on Genetic and evolutionary computation
An Information-Theoretic Analysis on the Interactions of Variables in Combinatorial Optimization Problems

Evolutionary Computation
A note on problem difficulty measures in black-box optimization: Classification, realizations and predictability

Evolutionary Computation
Cross-disciplinary perspectives on meta-learning for algorithm selection

ACM Computing Surveys (CSUR)
Information Theoretic Classification of Problems for Metaheuristics

SEAL '08 Proceedings of the 7th International Conference on Simulated Evolution and Learning
Empirical hardness models: Methodology and a case study on combinatorial auctions

Journal of the ACM (JACM)
Quantifying ruggedness of continuous landscapes using entropy

CEC'09 Proceedings of the Eleventh conference on Congress on Evolutionary Computation
A Knowledge Discovery Approach to Understanding Relationships between Scheduling Problem Structure and Heuristic Performance

Learning and Intelligent Optimization
Benchmarking evolutionary algorithms: towards exploratory landscape analysis

PPSN'10 Proceedings of the 11th international conference on Parallel problem solving from nature: Part I
Real-valued multimodal fitness landscape characterization for evolution

ICONIP'10 Proceedings of the 17th international conference on Neural information processing: theory and algorithms - Volume Part I
Global characterization of the CEC 2005 fitness landscapes using fitness-distance analysis

EvoApplications'11 Proceedings of the 2011 international conference on Applications of evolutionary computation - Volume Part I
Programming by optimization

Communications of the ACM
Performance prediction and automated tuning of randomized and parametric algorithms

CP'06 Proceedings of the 12th international conference on Principles and Practice of Constraint Programming
No free lunch theorems for optimization

IEEE Transactions on Evolutionary Computation
Design of evolutionary algorithms-A statistical perspective

IEEE Transactions on Evolutionary Computation

Recent advances in problem understanding: changes in the landscape a year on

Proceedings of the 15th annual conference companion on Genetic and evolutionary computation
On the use of meta-learning for instance selection: An architecture and an experimental study

Information Sciences: an International Journal

Quantified Score

Hi-index	0.00

Visualization

Abstract

Algorithm selection and configuration is a challenging problem in the continuous optimization domain. An approach to tackle this problem is to develop a model that links landscape analysis measures and algorithm parameters to performance. This model can be then used to predict algorithm performance when a new optimization problem is presented. In this paper, we investigate the use of a machine learning framework to build such a model. We demonstrate the effectiveness of our technique using CMA-ES as a representative algorithm and a feed-forward backpropagation neural network as the learning strategy. Our experimental results show that we can build sufficiently accurate predictions of an algorithm's expected performance. This information is used to rank the algorithm parameter settings based on the current problem instance, hence increasing the probability of selecting the best configuration for a new problem.