Algorithm Portfolios for Noisy Optimization: Compare Solvers Early

Cauwet, Marie-Liesse; Liu, Jialin; Teytaud, Olivier

doi:10.1007/978-3-319-09584-4_1

Cited by 13 publications

(6 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this paper, we apply AS to the black-box noisy optimization problem. This paper extends [12] with respect to (i) showing that the lag is necessary; (ii) extending experimental results; (iii) improving the convergence rates thanks to an unfair distribution of the computational budget.…”

Section: Introductionmentioning

confidence: 52%

“…Some AS algorithms [20,2] do not need a separate training phase, and perform entirely online solver selection; a weakness of this approach is that it is only possible when a large enough budget is available, so that the training phase has a minor cost. A portfolio algorithm, namely Noisy Optimization Portfolio Algorithm (NOPA), designed for noisy optimization solvers, and which distributes uniformly the computational power among them, is proposed in [12]. We extend it to INOPA (Improved NOPA), which is allowed to distribute the budget in an unfair manner.…”

Section: Algorithm Selectionmentioning

confidence: 99%

See 1 more Smart Citation

Algorithm portfolios for noisy optimization

Cauwet

Liu

Rozière

et al. 2015

Ann Math Artif Intell

Self Cite

View full text Add to dashboard Cite

Noisy optimization is the optimization of objective functions corrupted by noise. A portfolio of solvers is a set of solvers equipped with an algorithm selection tool for distributing the computational power among them. Portfolios are widely and successfully used in combinatorial optimization.In this work, we study portfolios of noisy optimization solvers. We obtain mathematically proved performance (in the sense that the portfolio performs nearly as well as the best of its solvers) by an ad hoc portfolio algorithm dedicated to noisy optimization. A somehow surprising result is that it is better to compare solvers with some lag, i.e., propose the current recommendation of best solver based on their performance earlier in the run. An additional finding is a principled method for distributing the computational power among solvers in the portfolio.

show abstract

Section: Introductionmentioning

confidence: 52%

Section: Algorithm Selectionmentioning

confidence: 99%

Algorithm portfolios for noisy optimization

Cauwet

Liu

Rozière

et al. 2015

Ann Math Artif Intell

Self Cite

View full text Add to dashboard Cite

show abstract

“…λ I represents the number of points selected (more generally termed individuals size) during one iteration, r n is the number of resamplings of each individual and σ n is the stepsize. We refer to [25] for more details on the settings and the different algorithms. As shown in Table I, 24 solvers are used to resolve the Cart-Pole problem, thus resulting in 24 options in the nonstationary bandit problem.…”

Section: A Non-stationary Datamentioning

confidence: 99%

Differential Evolution algorithm applied to non-stationary bandit problem

St-Pierre

Liu

2014

2014 IEEE Congress on Evolutionary Computation (CEC)

Self Cite

View full text Add to dashboard Cite

In this paper we compare Differential Evolution (DE), an evolutionary algorithm, to classical bandit algorithms over the non-stationary bandit problem. First we define a testcase where the variation of the distributions depends on the number of times an option is evaluated rather than over time. This definition allows the possibility to apply these algorithms over a wide range of problems such as black-box portfolio selection. Second we present our own variant of discounted Upper Confidence Bound (UCB) algorithm that outperforms the current state-of-the-art algorithms for the non-stationary bandit problem. Third, we introduce a variant of DE and show that, on a selection over a portfolio of solvers for the Cart-Pole problem, our version of DE outperforms the current best UCB algorithms.

show abstract

“…For instance, using a threshold (Rudolph, 2001) that is related to the noise characteristics to avoid making comparisons of individuals that might, in fact, be very similar or statistically the same; this is usually called threshold selection and can be applied either to explicit or implicit averaging fitness functions. The algorithms used for solution, themselves, can be also tested, with some authors proposing, instead of taking more measures, testing different solvers (Cauwet et al, 2014), some of which might be more affected by noise than others.…”

Section: State Of the Artmentioning

confidence: 99%

Studying and Tackling Noisy Fitness in Evolutionary Design of Game Characters

Merelo

Castillo

Mora

et al. 2014

Proceedings of the International Conference on Evolutionary Computation Theory and Applications

View full text Add to dashboard Cite

In most computer games as in life, the outcome of a match is uncertain due to several reasons: the characters or assets appear in different initial positions or the response of the player, even if programmed, is not deterministic; different matches will yield different scores. That is a problem when optimizing a game-playing engine: its fitness will be noisy, and if we use an evolutionary algorithm it will have to deal with it. This is not straightforward since there is an inherent uncertainty in the true value of the fitness of an individual, or rather whether one chromosome is better than another, thus making it preferable for selection. Several methods based on implicit or explicit average or changes in the selection of individuals for the next generation have been proposed in the past, but they involve a substantial redesign of the algorithm and the software used to solve the problem. In this paper we propose new methods based on incremental computation (memory-based) or fitness average or, additionally, using statistical tests to impose a partial order on the population; this partial order is considered to assign a fitness value to every individual which can be used straightforwardly in any selection function. Tests using several hard combinatorial optimization problems show that, despite an increased computation time with respect to the other methods, both memory-based methods have a higher success rate than implicit averaging methods that do not use memory; however, there is not a clear advantage in success rate or algorithmic terms of one method over the other.

show abstract

Algorithm Portfolios for Noisy Optimization: Compare Solvers Early

Cited by 13 publications

References 27 publications

Algorithm portfolios for noisy optimization

Algorithm portfolios for noisy optimization

Differential Evolution algorithm applied to non-stationary bandit problem

Studying and Tackling Noisy Fitness in Evolutionary Design of Game Characters

Contact Info

Product

Resources

About