Hyper-Reactive Tabu Search for MaxSAT

Ansótegui, Carlos; Heymann, Britta; Pon, Josep; Sellmann, Meinolf; Tierney, Kevin

doi:10.1007/978-3-030-05348-2_27

Cited by 9 publications

(7 citation statements)

References 18 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In this context, we refer only to the well-known handbook [2] and, in particular, to its chapter on MaxSAT [24]. It should be noted that in a number of papers various metaheurists were used to solve MaxSAT, employing both local search (see [25,26], etc.) and the concept of evolutionary computations (see, for example, [26,27]).…”

Section: Preliminary Computational Resultsmentioning

confidence: 99%

Merging Variables: One Technique of Search in Pseudo-Boolean Optimization

Семенов

2019

Mathematical Optimization Theory and Operations Research

View full text Add to dashboard Cite

In the present paper we describe new heuristic technique, which can be applied to the optimization of pseudo-Boolean functions including Black-Box functions. This technique is based on a simple procedure which consists in transition from the optimization problem over Boolean hypercube to the optimization problem of auxiliary function in a specially constructed metric space. It is shown that there is a natural connection between the points of the original Boolean hypercube and points from the new metric space. For the Boolean hypercube with fixed dimension it is possible to construct a number of such metric spaces. The proposed technique can be considered as a special case of Variable Neighborhood Search, which is focused on pseudo-Boolean optimization. Preliminary computational results show high efficiency of the proposed technique on some reasonably hard problems. Also it is shown that the described technique in combination with the well-known (1+1)-Evolutionary Algorithm allows to decrease the upper bound on the runtime of this algorithm for arbitrary pseudo-Boolean functions.

show abstract

Section: Preliminary Computational Resultsmentioning

confidence: 99%

Merging Variables: One Technique of Search in Pseudo-Boolean Optimization

Семенов

2019

Mathematical Optimization Theory and Operations Research

View full text Add to dashboard Cite

show abstract

“…provide the first formal definition of the DAC setting, however, there is a significant amount of earlier work for learning dynamic configuration policies Littman, 2000, 2001;Pettinger and Everson, 2002). Such earlier works label the setting as parameter control (Karafotias et al, 2015), online algorithm selection (Vermetten et al, 2019), adaptive selection/configuration (Fialho et al, 2010;van Rijn et al, 2018), or hyper-reactive search (Ansótegui et al, 2017(Ansótegui et al, , 2018a. For a comprehensive overview of parameter control with respect to parameters of evolutionary algorithms, we refer the interested reader to Karafotias et al (2015).…”

Section: Dynamic Methodsmentioning

confidence: 99%

“…Hyper-configurable reactive search Under the label hyper-configurable reactive search (HCRS), Ansótegui et al (2017), Ansótegui et al (2018a) and Sellmann and Tierney (2020) propose an approach to DAC by using standard, offline AC techniques to train a dynamic policy. More specifically, instead of tuning the parameters of an algorithm directly, each parameter is determined by a logistic regression that accepts runtime features from the target algorithm.…”

Section: Dynamic Methodsmentioning

confidence: 99%

A Survey of Methods for Automated Algorithm Configuration

Schede¹,

Brandt²,

Tornede³

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Algorithm configuration (AC) is concerned with the automated search of the most suitable parameter configuration of a parametrized algorithm. There is currently a wide variety of AC problem variants and methods proposed in the literature. Existing reviews do not take into account all derivatives of the AC problem, nor do they offer a complete classification scheme. To this end, we introduce taxonomies to describe the AC problem and features of configuration methods, respectively. We review existing AC literature within the lens of our taxonomies, outline relevant design choices of configuration approaches, contrast methods and problem variants against each other, and describe the state of AC in industry. Finally, our review provides researchers and practitioners with a look at future research directions in the field of AC.

show abstract

“…Finally, we need to determine the feature weights w ∈ R n to fully define the scoring function. We use an algorithm configurator to determine the hyperparameters, as was previously proposed in [Ansótegui et al, 2017[Ansótegui et al, , 2018, in which AC is used for determining weights of linear regressions in a reactive search. We tune on a set of 43 black box optimization problems from the 2160 problems in Hansen et al [2020].…”

Section: Hyperparametersmentioning

confidence: 99%

“…The hyperparameters of our approach control a scoring function that guides the creation of batches of candidates, and we select our final batch of candidates based on which candidates appear most often in the sampled batches. It turns out that setting the hyperparameters of our heuristic is itself a type of black-box optimization (BBO) problem that we solve using the algorithm configurator GGA [Ansotegui et al, 2015] as in Ansótegui et al [2017] and Ansótegui et al [2018]. In this way, our heuristic can be customized to a specific domain, although, in this work, we try to make it work generally for BBO functions.…”

Section: Introductionmentioning

confidence: 99%

Learning to Optimize Black-Box Functions with Extreme Limits on the Number of Function Evaluations

Ansótegui

Sellmann²,

Shah

et al. 2021

Lecture Notes in Computer Science

Self Cite

View full text Add to dashboard Cite

We consider black-box optimization in which only an extremely limited number of function evaluations, on the order of around 100, are affordable and the function evaluations must be performed in even fewer batches of a limited number of parallel trials. This is a typical scenario when optimizing variable settings that are very costly to evaluate, for example in the context of simulation-based optimization or machine learning hyperparameterization. We propose an original method that uses established approaches to propose a set of points for each batch and then down-selects from these candidate points to the number of trials that can be run in parallel. The key novelty of our approach lies in the introduction of a hyperparameterized method for down-selecting the number of candidates to the allowed batch-size, which is optimized offline using automated algorithm configuration. We tune this method for black box optimization and then evaluate on classical black box optimization benchmarks. Our results show that it is possible to learn how to combine evaluation points suggested by highly diverse black box optimization methods conditioned on the progress of the optimization. Compared with the state of the art in black box minimization and various other methods specifically geared towards few-shot minimization, we achieve an average reduction of 50% of normalized cost, which is a highly significant improvement in performance.

show abstract

Hyper-Reactive Tabu Search for MaxSAT

Cited by 9 publications

References 18 publications

Merging Variables: One Technique of Search in Pseudo-Boolean Optimization

Merging Variables: One Technique of Search in Pseudo-Boolean Optimization

A Survey of Methods for Automated Algorithm Configuration

Learning to Optimize Black-Box Functions with Extreme Limits on the Number of Function Evaluations

Contact Info

Product

Resources

About