The oracle penalty method

Schlüter, Martin; Gerdts, Matthias

doi:10.1007/s10898-009-9477-0

Cited by 101 publications

(64 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Stability studies of the solutions have been performed, which show this optimization method provides most often with local maxima. Better rejection rates may be obtained, for example, by using a global optimization based on the Ant algorithm [48]. We tested it on a a small subset of variables, which yielded promising results.…”

Section: Nimentioning

confidence: 99%

Taming Random Lasers through Active Spatial Control of the Pump

et al. 2012

View full text Add to dashboard Cite

Active control of the pump spatial profile is proposed to exercise control over random laser emission. We demonstrate numerically the selection of any desired lasing mode from the emission spectrum. An iterative optimization method is employed, first in the regime of strong scattering where modes are spatially localized and can be easily selected using local pumping. Remarkably, this method works efficiently even in the weakly scattering regime, where strong spatial overlap of the modes precludes spatial selectivity. A complex optimized pump profile is found, which selects the desired lasing mode at the expense of others, thus demonstrating the potential of pump shaping for robust and controllable singlemode operation of a random laser.PACS numbers: 42.55. Zz,42.25.Dd Multiple scattering of light in random media can be actively manipulated through spatial shaping of the incident wavefront [1]. This technique has allowed advances in focusing [2][3][4], and imaging [5,6], paving the road to actual control of light transport in strongly scattering media [7][8][9]. Introducing gain in disordered media allows amplification of multiply scattered light, leading to the observation of random lasing [10]. The broad range of systems where it has been studied [11] and the fundamental questions it has raised [12,13] has captivated the community for this last decade. Prospective of random lasing is however strongly hindered by the absence of emission control: random lasers are highly multimode with unpredictable lasing frequencies and polydirectional output. Manipulation of the underlying random structure [14][15][16][17][18][19][20][21][22][23] and recent work constraining the range of lasing frequencies [24,25] have resulted in significant progress toward possible control. However, the ability to choose a specific frequency in generic random lasing systems has not yet been achieved. The spatial profile of the pump is an interesting degree of freedom readily available in random laser (e.g., [26,27]). In a regime of very strong scattering where the modes of the random system are spatially localized [28], local pumping allows selection of spatially non-overlapping modes [29,30]. In weaker scattering media however (e.g., [31][32][33]), several hurdles appear toward achieving fine control. Selecting modes is complicated by a narrow distribution of lasing thresholds [34,35] and spatial mode overlap. Increased pumping required in these lossy systems begins to alter the random laser itself. Moreover, modifying the shape of the pump introduces changes to both the spatial and spectral properties of lasing modes [33,[36][37][38]. Such difficulties are typically absent in more conventional lasers, which have employed pump shaping, both electrically [39][40][41][42] and optically [43], to select favorable lasing modes. The question is, can shaping of the incident pump field achieve taming of random lasers?In this letter, we exercise control over the distribution of lasing thresholds via the pump geometry to choose the random laser em...

show abstract

Section: Nimentioning

confidence: 99%

Taming Random Lasers through Active Spatial Control of the Pump

et al. 2012

View full text Add to dashboard Cite

show abstract

“…The evolutionary algorithm within MIDACO is based on the ant colony optimization metaheuristic for continuous search domains proposed by Socha and Dorigo [26] and was extended to mixed-integer domains by Schlueter et al in [17]. For constrained optimization problems the algorithm applies the Oracle Penalty Method which was introduced in Schlueter and Gerdts [18]. While the MIDACO algorithm is conceptually designed as general black-box solver, it has proven its effectiveness especially on challenging interplanetary space trajectory design problems (see Schlueter [21]), where it holds several best known record solutions on benchmarks provided by the European Space Agency [5].…”

Section: Midaco Algorithmmentioning

confidence: 99%

MIDACO Parallelization Scalability on 200 MINLP Benchmarks

Schlueter

Munetomo

2017

Journal of Artificial Intelligence and Soft Computing Research

View full text Add to dashboard Cite

This contribution presents a numerical evaluation of the impact of parallelization on the performance of an evolutionary algorithm for mixed-integer nonlinear programming (MINLP). On a set of 200 MINLP benchmarks the performance of the MIDACO solver is assessed with gradually increasing parallelization factor from one to three hundred. The results demonstrate that the efficiency of the algorithm can be significantly improved by parallelized function evaluation. Furthermore, the results indicate that the scale-up behaviour on the efficiency resembles a linear nature, which implies that this approach will even be promising for very large parallelization factors. The presented research is especially relevant to CPU-time consuming real-world applications, where only a low number of serial processed function evaluation can be calculated in reasonable time.

show abstract

“…To treat four constraints efficiently, we employ the oracle penalty method [10] in which penalty is dynamically controlled. This method allows us to avoid the situation that the penalty becomes overwhelmingly dominant in F. …”

Section: A Problem Settingmentioning

confidence: 99%

Multimaterial Topology Optimization of Electric Machines Based on Normalized Gaussian Network

2015

View full text Add to dashboard Cite

This paper presents a topology optimization method for multimaterial models based on the normalized Gaussian network (NGnet). In this method, one can determine optimal shapes of machines which are composed of various materials such as iron, magnet and nonmagnetic material. The present method is applied to the optimization of interior permanent magnet motor to determine the distributions of magnetic core and flux barrier as well as magnets. The optimization results show that average torque can be improved using as small amount of magnet as possible. In addition, characteristic of the present method is discussed in detail.Index Terms-interior permanent magnet motor, topology optimization.

show abstract

The oracle penalty method

Abstract: Constrained optimization, Global optimization, Penalty function, Stochastic metaheuristic, Ant colony optimization, MIDACO-Solver, Mixed integer nonlinear programming (MINLP),

Cited by 101 publications

References 22 publications

Taming Random Lasers through Active Spatial Control of the Pump

Taming Random Lasers through Active Spatial Control of the Pump

MIDACO Parallelization Scalability on 200 MINLP Benchmarks

Multimaterial Topology Optimization of Electric Machines Based on Normalized Gaussian Network

Contact Info

Product

Resources

About