Abstract
Constraints in optimization come traditionally in two types familiar to most readers: hard and soft. Hard constraints delineate absolutely between feasible and infeasible solutions, whereas soft constraints essentially specifyadditional objectives. In this chapter, we describe a third type of constraint, much less familiar and only investigated recently, which we call ephemeral resource constraints (ERCs). ERCs differ from the other constraints in three major ways. (i) The constraints are dynamic or temporary (i.e., may be active or not active), and occur only during optimization—they do not affect the feasibility of final solutions. (ii) Solutions violating the constraints cannot be evaluated on the objective function—in fact that is their main defining property. (iii) The constraints that are active are usually a function of previous solutions evaluated, bringing in a time-linkage aspect to the optimization. We explain with examples how these constraints arise in real-world optimization problems, especially when solution evaluation depends on experimental processes (i.e. in “closed-loop optimization”). Using a theoretical model based on Markov chains, the effects of these constraints on evolutionary search, e.g., drift effects on the search direction, are described. Next, a number of strategies for coping with ERCs are summarized, and evidence for their robustness is provided. In the final section, we look to the future and consider the many open questions there are in this new area.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Notes
- 1.
When an EA is used, closed-loop optimization may also be referred to as evolutionary experimentation (Rechenberg 2000) or experimental evolution.
- 2.
Indeed, we can consider any optimization problem or benchmark.
- 3.
We leave out the variables \(t_{\text {ctf}}^{\text {start}},t_{\text {ctf}}^{\text {end}}, c_{\text {order}} ,c_{\mathrm{{time\_step}}},\) and \(C\) from \( commCompERC (\ldots )\) for ease of presentation. They will be specified where appropriate.
- 4.
Note, in an EA performing optimization of a function, the number of performed selection steps displayed on the x-axes of Fig. 4.6 would be equivalent to the number of performed function evaluations.
- 5.
We get the zigzag-shaped line for SSGA (rri) during the constraint time frame because \(c_t(B)\) is plotted after each time step containing here of one selection step. For GGA the change in \(c_t(B)\) is smooth because a time step consists of \(\mu \) selection steps.
- 6.
For D-MAB we set the threshold parameter to \(\lambda _{\text {PH}}=0.1\), the tolerance parameter to \(\delta =0.01\), and the scaling factor to \(C=1\).
- 7.
RL-EA also employed the \(\epsilon \)-greedy action selection method (\(\epsilon =0.1\)), optimistic initial values for the action-value estimates, and replacing eligibility traces with the eligibility trace being set to 0 at the beginning of each EA run. The decay factor was set to \(\lambda =1\), the discount factor to \(\gamma =1\), and the learning rate to \(\alpha =0.1\).
- 8.
The instance considered is a uniform random 3-SAT problem and can be downloaded online at http://people.cs.ubc.ca/~hoos/SATLIB/benchm.html; the name of the instance is “uf50-218/uf50-01.cnf”. The instance consists of 218 clauses and is satisfiable. We treat this 3-SAT instance as a MAX-SAT optimization problem, with fitness calculated as the proportion of satisfied clauses.
References
Allmendinger R (2012) Tuning evolutionary search for closed-loop optimization. PhD thesis, Department of Computer Science, University of Manchester, UK
Allmendinger R, Knowles J (2010) On-line purchasing strategies for an evolutionary algorithm performing resource-constrained optimization. In: Proceedings of parallel problem solving from nature, pp 161–170
Allmendinger R, Knowles J (2011) Policy learning in resource-constrained optimization. In: Proceedings of the genetic and evolutionary computation conference, pp 1971–1978
Allmendinger R, Knowles J (2013) On handling ephemeral resource constraints in evolutionary search. Evol Comput 21(3):497–531
Auger A, Doerr B (2011) Theory of randomized search heuristics. World Scientific, Singapore
Bäck T, Knowles J, Shir OM (2010) Experimental optimization by evolutionary algorithms. In: Proceedings of the genetic and evolutionary computation conference (companion), pp 2897–2916
Bedau MA (2010) Coping with complexity: machine learning optimization of highly synergistic biological and biochemical systems. In: Keynote talk at the international conference on genetic and evolutionary computation
Borodin A, El-Yaniv R (1998) Online computation and competitive analysis. Cambridge University Press, Cambridge
Bosman PAN (2005) Learning, anticipation and time-deception in evolutionary online dynamic optimization. In: Proceedings of genetic and evolutionary computation conference, pp 39–47
Bosman PAN, Poutré HL (2007) Learning and anticipation in online dynamic optimization with evolutionary algorithms: the stochastic case. In: Proceedings of genetic and evolutionary computation conference, pp 1165–1172
Branke J (2001) Evolutionary optimization in dynamic environments. Kluwer Academic Publishers, Dordrecht
Caschera F, Gazzola G, Bedau MA, Moreno CB, Buchanan A, Cawse J, Packard N, Hanczyc MM (2010) Automated discovery of novel drug formulations using predictive iterated high throughput experimentation. PLoS ONE 5(1):e8546
Chen T, He J, Sun G, Chen G, Yao X (2009) A new approach for analyzing average time complexity of population-based evolutionary algorithms on unimodal problems. IEEE Trans Syst Man Cybern B 39(5):1092–1106
Coello CAC (2002) Theoretical and numerical constraint-handling techniques used with evolutionary algorithms: a survey of the state of the art. Comput Methods Appl Mech Eng 191(11–12):1245–1287
Costa LD, Fialho A, Schoenauer M, Sebag M (2008) Adaptive operator selection with dynamic multi-armed bandits. In: Proceedings of genetic and evolutionary computation conference, pp 913–920
Davis TE, Principe JC (1993) A Markov chain framework for the simple genetic algorithm. Evol Comput 1(3):269–288
Doob JL (1953) Stochastic processes. Wiley, New York
Finkel DE, Kelley CT (2009) Convergence analysis of sampling methods for perturbed Lipschitz functions. Pac J Optim 5:339–350
Goldberg DE, Segrest P (1987) Finite Markov chain analysis of genetic algorithms. In: Proceedings of the international conference on genetic algorithms, pp 1–8
Hartland C, Gelly S, Baskiotis N, Teytaud O, Sebag M (2006) Multi-armed bandits, dynamic environments and meta-bandits. In: NIPS workshop online trading of exploration and exploitation
Hartland C, Baskiotis N, Gelly S, Sebag M, Teytaud O (2007) Change point detection and meta-bandits for online learning in dynamic environments. In: CAp, pp 237–250
He J, Yao X (2002) From an individual to a population: an analysis of the first hitting time of population-based evolutionary algorithms. IEEE Trans Evol Comput 6(5):495–511
Herdy M (1997) Evolutionary optimization based on subjective selection-evolving blends of coffee. In: European congress on intelligent techniques and soft computing, pp 640–644
Holland JH (1975) Adaptation in natural and artificial systems. MIT Press, Boston
Horn J (1993) Finite Markov chain analysis of genetic algorithms with niching. In: Proceedings of the international conference on genetic algorithms, pp 110–117
Jin Y (2011) Surrogate-assisted evolutionary computation: recent advances and future challenges. Swarm Evol Comput 1(2):61–70
Judson RS, Rabitz H (1992) Teaching lasers to control molecules. Phys Rev Lett 68(10):1500–1503
Kauffman S (1989) Adaptation on rugged fitness landscapes. In: Lecture notes in the sciences of complexity, pp 527–618
Kaufman L, Rousseeuw PJ (1990) Finding groups in data: an introduction to cluster analysis. Wiley, New York
King RD, Whelan KE, Jones FM, Reiser PGK, Bryant CH, Muggleton SH, Kell DB, Oliver SG (2004) Functional genomic hypothesis generation and experimentation by a robot scientist. Nature 427:247–252
Klockgether J, Schwefel H-P (1970) Two-phase nozzle and hollow core jet experiments. In: Engineering aspects of magnetohydrodynamics, pp 141–148
Knowles J (2009) Closed-loop evolutionary multiobjective optimization. IEEE Comput Intell Mag 4(3):77–91
Lehre PK (2011) Fitness-levels for non-elitist populations. In: Proceedings of the conference on genetic and evolutionary computation, pp 2075–2082
Liepins GE, Potter WD (1991) A genetic algorithm approach to multiple-fault diagnosis. In: Handbook of genetic algorithms, pp 237–250
Mahfoud SW (1991) Finite Markov chain models of an alternative selection strategy for the genetic algorithm. Complex Syst 7:155–170
Michalewicz Z, Schoenauer M (1996) Evolutionary algorithms for constrained parameter optimization problems. Evol Comput 4(1):1–32
Nakama T (2008) Theoretical analysis of genetic algorithms in noisy environments based on a Markov model. In: Proceedings of the genetic and evolutionary computation conference, pp 1001–1008
Nguyen TT (2010) Continuous dynamic optimisation using evolutionary algorithms. PhD thesis, University of Birmingham
Nix A, Vose MD (1992) Modeling genetic algorithms with Markov chains. Ann Math Artif Intell 5:79–88
Nocedal J, Wright SJ (1999) Numerical optimization. Springer, New York
Norris JR (1998) Markov chains (Cambridge Series in Statistical and Probabilistic Mathematics). Cambridge University Press, Cambridge
O’Hagan S, Dunn WB, Brown M, Knowles J, Kell DB (2005) Closed-loop, multiobjective optimization of analytical instrumentation: gas chromatography/time-of-flight mass spectrometry of the metabolomes of human serum and of yeast fermentations. Anal Chem 77(1):290–303
O’Hagan S, Dunn WB, Knowles J, Broadhurst D, Williams R, Ashworth JJ, Cameron M, Kell DB (2007) Closed-loop, multiobjective optimization of two-dimensional gas chromatography/mass spectrometry for serum metabolomics. Anal Chem 79(2):464–476
Pettinger JE, Everson RM (2003) Controlling genetic algorithms with reinforcement learning. Technical report, The University of Exeter
Rechenberg I (2000) Case studies in evolutionary experimentation and computation. Comput Methods Appl Mech Eng 2–4(186):125–140
Reeves CR, Rowe JE (2003) Genetic algorithms—principles and perspectives: a guide to GA theory. Kluwer Academic Publishers, Boston
Rummery GA, Niranjan M (1994) On-line Q-learning using connectionist systems. Technical report CUED/F-INFENG/TR 166, Cambridge University Engineering Department
Schwefel H-P (1968) Experimentelle Optimierung einer Zweiphasendüse, Teil 1. AEG Research Institute Project MHD-Staustrahlrohr 11.034/68, Technical report 35, Berlin
Schwefel H-P (1975) Evolutionsstrategie und numerische Optimierung. PhD thesis, Technical University of Berlin
Shir O, Bäck T (2009) Experimental optimization by evolutionary algorithms. In: Tutorial at the international conference on genetic and evolutionary computation
Shir OM (2008) Niching in derandomized evolution strategies and its applications in quantum control: a journey from organic diversity to conceptual quantum designs. PhD thesis, University of Leiden
Small BG, McColl BW, Allmendinger R, Pahle J, López-Castejón G, Rothwell NJ, Knowles J, Mendes P, Brough D, Kell DB (2011) Efficient discovery of anti-inflammatory small molecule combinations using evolutionary computing. Nat Chem Biol (to appear)
Sutton RS, Barto AG (1998) Reinforcement learning: an introduction. MIT Press, Cambridge
Syswerda G (1989) Uniform crossover in genetic algorithms. In: Proceedings of the international conference on genetic algorithms, pp 2–9
Syswerda G (1991) A study of reproduction in generational and steady state genetic algorithms. In: Foundations of genetic algorithms, pp 94–101
Thompson A (1996) Hardware evolution: automatic design of electronic circuits in reconfigurable hardware by artificial evolution. PhD thesis, University of Sussex
Vaidyanathan S, Broadhurst DI, Kell DB, Goodacre R (2003) Explanatory optimization of protein mass spectrometry via genetic search. Anal Chem 75(23):6679–6686
Vose MD, Liepins GE (1991) Punctuated equilibria in genetic search. Complex Syst 5:31–44
Zhang W (2001) Phase transitions and backbones of 3-SAT and maximum 3-SAT. In: Proceedings of the international conference on principles and practice of constraint programming, pp 153–167
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2015 Springer India
About this chapter
Cite this chapter
Allmendinger, R., Knowles, J. (2015). Ephemeral Resource Constraints in Optimization. In: Datta, R., Deb, K. (eds) Evolutionary Constrained Optimization. Infosys Science Foundation Series(). Springer, New Delhi. https://doi.org/10.1007/978-81-322-2184-5_4
Download citation
DOI: https://doi.org/10.1007/978-81-322-2184-5_4
Published:
Publisher Name: Springer, New Delhi
Print ISBN: 978-81-322-2183-8
Online ISBN: 978-81-322-2184-5
eBook Packages: EngineeringEngineering (R0)