Abstract
The No Free Lunch (NFL) theorems for search and optimization are reviewed and their implications for the design of metaheuristics are discussed. The theorems state that any two search or optimization algorithms are equivalent when their performance is averaged across all possible problems and even over subsets of problems fulfilling certain constraints. The NFL results show that if there is no assumption regarding the relation between visited and unseen search points, efficient search and optimization is impossible. There is no well-performing universal metaheuristic, but the heuristics must be tailored to the problem class at hand using prior knowledge. In practice, it is not likely that the preconditions of the NFL theorems are fulfilled for a problem class and thus differences between algorithms exist. Therefore, tailored algorithms can exploit structure underlying the optimization problem. Given full knowledge about the problem class, it is, in theory, possible to construct an optimal algorithm.
Keywords
- Metaheuristics
- Search Points
- Restricted Problem Class
- Optimal Search Algorithm
- Finite Horizon Optimal Control Problem
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
Thus not only our reason fails us in the discovery of the ultimate connexion of causes and effects, but even after experience has informed us of their constant conjunction, it is impossible for us to satisfy ourselves by our reason, why we should extend that experience beyond those particular instances, which have fallen under our observation. We suppose, but are never able to prove, that there must be a resemblance betwixt those objects, of which we have had experience, and those which lie beyond the reach of our discovery. (David Hume, 1739 [10, 14])
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
In general, this cannot be done in the theorems because \(\sum _{f\in F}\delta (k,\sum _{a\prime\in A}p_{a}(a\prime)c(Y (f,m,a\prime)))\neq \sum _{f\in F}\sum _{a\prime\in A}p_{a}(a\prime)\delta (k,c(Y (f,m,a\prime)))\).
- 2.
It is also possible to map the optimization to an infinite horizon problem with appropriate absorbing states.
- 3.
In a S-MDP no state is ever revisited. Hence, for any policy, the transition probability graph is acyclic and thus value iteration finds the optimal policy after at most m steps, where each step needs \(O{(\vert \mathcal{A}\vert \vert \mathcal{S}\vert )}^{2}\) computations (see [3, Sect. 2.2.2] or [2]).
References
A. Auger, O. Teytaud, Continuous lunches are free plus the design of optimal optimization algorithms. Algorithmica 57(1), 121–146 (2010)
D.P. Bertsekas, Dynamic programming and optimal control. Athena Sci. 2 (2007)
D.P. Bertsekas, J.N. Tsitsiklis, Neuro-dynamic programming. Athena Sci. (1996)
O. Bousquet, S. Boucheron, G. Lugosi, Introduction to statistical learning theory, in Advanced Lectures in Machine Learning, ed. by O. Bousquet, U. von Luxburg, G. Rätsch. LNAI, vol. 3176 (Springer, Berlin Heidelberg, 2004), pp. 169–207
D.W. Corne, J.D. Knowles, No free lunch and free leftovers theorems for multiobjective optimisation problems, in Evolutionary Multi-Criterion Optimization (EMO 2003), ed. by C.M. Fonseca, P.J. Fleming, E. Zitzler, L. Thiele, K. Deb. LNCS, vol. 2632 (Springer, Berlin Heidelberg, 2003), pp. 327–341
S. Droste, T. Jansen, I. Wegener, Optimization with randomized search heuristics – the (A)NFL theorem, realistic scenarios, and difficult functions. Theor. Comput. Sci. 287(1), 131–144 (2002)
T.M. English, Evaluation of evolutionary and genetic optimizers: no free lunch, in Proceedings of the Fifth Annual Conference on Evolutionary Programming (EP V), San Diego, ed. by L.J. Fogel, P.J. Angeline, T. Bäck (MIT, 1996), pp. 163–169
T.M. English, Optimization is easy and learning is hard in the typical function, in Proceedings of the 2000 Congress on Evolutionary Computation (CEC 2000), San Diego, ed. by A. Zalzala, C. Fonseca, J.H. Kim, A. Smith (IEEE, 2000), pp. 924–931
T. English, On the structure of sequential search: beyond “No free lunch”, in Evolutionary Computation in Combinatorial Optimization (EvoCOP 2004), ed. by J. Gottlieb, G.R. Raidl. LNCS, vol. 3004 (Springer, Berlin Heidelberg, 2004), pp. 95–103
C. Giraud-Carrier, F. Provost, Toward a justification of meta-learning: is the no free lunch theorem a show-stopper? in Proceedings of the ICML-2005 Workshop on Meta-learning, Bonn, 2005
N. Hansen: Invariance, self-adaptation and correlated mutations and evolution strategies, in Parallel Problem Solving from Nature (PPSN VI), ed. by M. Schoenauer, K. Deb, G. Rudolph, X. Yao, E. Lutton, J.J.M. Guervós, H.P. Schwefel. LNCS, vol. 1917 (Springer, Berlin Heidelberg, 2000), pp. 355–364
N. Hansen, Adaptive encoding: how to render search coordinate system invariant, in Parallel Problem Solving from Nature (PPSN X), ed. by G. Rudolph, T. Jansen, S.M. Lucas, C. Poloni, N. Beume. LNCS, vol. 5199 (Springer, Berlin Heidelberg, 2008), pp. 205–214
N. Hansen, A. Ostermeier, Completely derandomized self-adaptation in evolution strategies. Evol. Comput. 9(2), 159–195 (2001)
D. Hume, A Treatise of Human Nature, Chap. Sect. vi. Of the Inference from the impression to the idea. Web edition published by eBooks@Adelaide, 2006 (1739)
C. Igel, Recent results on no-free-lunch for optimization, in Theory of Evolutionary Algorithms, no. 04081, ed. by H. Beyer, T. Jansen, C. Reeves, M.D. Vose, in Dagstuhl Seminar Proceedings, Abstract Collection, Schloss Dagstuhl (Internationales Begegnungs- und Forschungszentrum für Informatik (IBFI), 2004)
C. Igel, P. Stagge, Graph isomorphisms effect structure optimization of neural networks, in International Joint Conference on Neural Networks (IJCNN 2002), Honolulu (IEEE, 2002), pp. 142–147
C. Igel, M. Toussaint, Neutrality and self-adaptation. Nat. Comput. 2(2), 117–132 (2003)
C. Igel, M. Toussaint, On classes of functions for which no free lunch results hold. Inf. Process. Lett. 86(6), 317–321 (2003)
C. Igel, M. Toussaint, Recent results on no-free-lunch theorems for optimization. arXiv preprint cs.NE/0303032 (2003), http://arxiv.org/abs/cs.NE/0303032
C. Igel, M. Toussaint, A no-free-lunch theorem for non-uniform distributions of target functions. J. Math. Model. Algorithms 3(4), 313–322 (2004)
M. Köppen, D.H. Wolpert, W.G. Macready, Remarks on a recent paper on the “no free lunch” theorems. IEEE Trans. Evol. Comput. 5(3), 295–296 (1995)
R. Motwani, P. Raghavan, Randomized Algorithms (Cambridge University Press, Cambridge, 1995)
J. Niehaus, C. Igel, W. Banzhaf, Reducing the number of fitness evaluations in graph genetic programming using a canonical graph indexed database. Evol. Comput. 15(2), 199–221 (2007)
N.J. Radcliffe, P.D. Surry, Fundamental limitations on search algorithms: evolutionary computing in perspective, in Computer Science Today: Recent Trends and Development, ed. by J. van Leeuwen. LNCS, vol. 1000 (Springer, Berlin Heidelberg, 1995), pp. 275–291
G.J.E. Rawlins, Introduction, in Foundations of Genetic Algorithms (FOGA), ed. by G.J.E. Rawlins (Morgan Kaufmann, San Mateo, 1991), pp. 1–10
J.E. Rowe, M.D. Vose, A.H. Wright, Reinterpreting no free lunch. Evol. Comput. 17(1), 117–129 (2009)
C. Schumacher, Fundamental limitations of search. Ph.D. Thesis, University of Tennessee, 2000
C. Schumacher, M.D. Vose, L.D. Whitley, The no free lunch and description length, in Proceedings of the Genetic and Evolutionary Computation Conference (GECCO 2001), San Francisco, ed. by L. Spector, E. Goodman, A. Wu, W. Langdon, H.M. Voigt, M. Gen, S. Sen, M. Dorigo, S. Pezeshk, M. Garzon, E. Burke (Morgan Kaufmann, 2001), pp. 565–570
M.J. Streeter, Two broad classes of functions for which a no free lunch result does not hold, in Proceedings of the Genetic and Evolutionary Computation Conference (GECCO 2003), Chicago, ed. by E. Cantú-Paz, J.A. Foster, K. Deb, D. Davis, R. Roy, U.M. O’Reilly, H.G. Beyer, R. Standish, G. Kendall, S. Wilson, M. Harman, J. Wegener, D. Dasgupta, M.A. Potter, A.C. Schultz, K. Dowsland, N. Jonoska, J. Miller. LNCS, vol. 2724 (Springer, Berlin Heidelberg, 2003), pp. 1418–1430
T. Suttorp, N. Hansen, C. Igel, Efficient covariance matrix update for variable metric evolution strategies. Mach. Learn. 75(2), 167–197 (2009). doi:10.1007/s10994-009-5102-1
D. Whitley, A free lunch proof for Gray versus binary encodings, in Proceedings of the Genetic and Evolutionary Computation Conference (GECCO 1999), Orlando, ed. by W. Banzhaf, J. Daida, A.E. Eiben, M.H. Garzon, V. Honavar, M. Jakiela, R.E. Smith, vol. 1 (Morgan Kaufmann, 1999), pp. 726–733
D. Whitley, J. Rowe, Focused no free lunch theorems, in Proceedings of the Genetic and Evolutionary Computation Conference (GECCO 2008), Atlanta, ed. by M. Keijzer, et al. (ACM, 2008), pp. 811–818
D. Whitley, J.P. Watson, Complexity theory and the no free lunch theorem, in Search Methodologies: Introductory Tutorials in Optimization and Decision Support Techniques, ed. by E.K. Burke, G. Kendall, Chap. 11 (Springer, New York, 2005), pp. 317–339
D.H. Wolpert, The lack of a priori distinctions between learning algorithms. Neural Comput. 8(7), 1341–1390 (1996)
D.H. Wolpert, W.G. Macready, No free lunch theorems for search. Technical Report SFI-TR-05-010, Santa Fe Institute, Santa Fe (1995)
D.H. Wolpert, W.G. Macready, No free lunch theorems for optimization. IEEE Trans. Evol. Comput. 1(1), 67–82 (1997)
D.H. Wolpert, W.G. Macready, Coevolutionary free lunches. IEEE Trans. Evol. Comput. 9(6), 721–735 (2005)
Acknowledgements
Parts of this text are based on former joint work with Marc Toussaint. The author acknowledges support from the German Federal Ministry of Education and Research within the National Network Computational Neuroscience – Bernstein Fokus: “Learning behavioral models: From human experiment to technical assistance”, grant FKZ 01GQ0951.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2014 Springer-Verlag Berlin Heidelberg
About this chapter
Cite this chapter
Igel, C. (2014). No Free Lunch Theorems: Limitations and Perspectives of Metaheuristics. In: Borenstein, Y., Moraglio, A. (eds) Theory and Principled Methods for the Design of Metaheuristics. Natural Computing Series. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-33206-7_1
Download citation
DOI: https://doi.org/10.1007/978-3-642-33206-7_1
Published:
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-33205-0
Online ISBN: 978-3-642-33206-7
eBook Packages: Computer ScienceComputer Science (R0)