Black Box Optimization Benchmarking of the GLOBAL Method

Pál, László; Csendes, Tibor; Markót, Mihály Csaba; Neumaier, Arnold

doi:10.1162/evco_a_00089

Cited by 16 publications

(9 citation statements)

References 23 publications

(22 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We compare through empirical studies our method with a variety of alternative multistart methods including MLSL, MATLAB's MultiStart, MATLAB's GlobalSearch, and GLOBAL when the number of allowable function evaluations is limited. This contrasts with some earlier results including those reported in Pál et al (2012), in which the number of function evaluations was set to as high as 2 × 10 4 × d. 3. In Section 3.4, we present a generic approach that allows us to create a test function that mimics the multimodal nature of objective functions arising in many black-box simulations (e.g., parameter calibration in simulation model ).…”

Section: Introductioncontrasting

confidence: 82%

“…Therefore, the number of sample points of GlobalSearch is expected to be larger than that of other methods. This was also pointed out by Pál et al (2012). The two main algorithm parameters of GlobalSearch are the number of trial points (N 1 ) and the number of points used in Stage 1 of the algorithm (N 2 ).…”

Section: Appendix B: Algorithm Parameters Used In Numerical Experimentsmentioning

confidence: 81%

See 1 more Smart Citation

SOMS: SurrOgate MultiStart algorithm for use with nonlinear programming for global optimization

Krityakierne

Shoemaker

2015

Int Trans Operational Res

View full text Add to dashboard Cite

SOMS is a general surrogate-based multistart algorithm, which is used in combination with any local optimizer to find global optima for computationally expensive functions with multiple local minima. SOMS differs from previous multistart methods in that a surrogate approximation is used by the multistart algorithm to help reduce the number of function evaluations necessary to identify the most promising points from which to start each nonlinear programming local search. SOMS's numerical results are compared with four well-known methods, namely, Multi-Level Single Linkage (MLSL), MATLAB's MultiStart, MATLAB's GlobalSearch, and GLOBAL. In addition, we propose a class of wavy test functions that mimic the wavy nature of objective functions arising in many black-box simulations. Extensive comparisons of algorithms on the wavy test functions and on earlier standard global-optimization test functions are done for a total of 19 different test problems. The numerical results indicate that SOMS performs favorably in comparison to alternative methods and does especially well on wavy functions when the number of function evaluations allowed is limited.

show abstract

Section: Introductioncontrasting

confidence: 82%

Section: Appendix B: Algorithm Parameters Used In Numerical Experimentsmentioning

confidence: 81%

SOMS: SurrOgate MultiStart algorithm for use with nonlinear programming for global optimization

Krityakierne

Shoemaker

2015

Int Trans Operational Res

View full text Add to dashboard Cite

show abstract

“…On the whole testbed we use the MATLAB's fmincon local search method in all dimensions. fmincon is an interior-point algorithm for constrained nonlinear problems which approximates the gradient using the finite difference method and based on a recent study [9], it performed well on most of the test functions. The maximum number of function evaluations for local search was set to 10% of the total budget, while the termination tolerance parameter value was set to 10 −12 .…”

Section: Methodsmentioning

confidence: 99%

Benchmarking a hybrid multi level single linkagealgorithm on the bbob noiseless testbed

Pál

2013

Proceedings of the 15th Annual Conference Companion on Genetic and Evolutionary Computation

Self Cite

View full text Add to dashboard Cite

Multi Level Single Linkage (MLSL) is a well known stochastic global optimization method. In this paper, a new hybrid variant (HMLSL) of the MLSL algorithm is presented. The most important improvements are related to the sampling phase: the sample is generated from a Sobol quasi-random sequence and a few percent of the population is further improved by using crossover and mutation operators like in a traditional differential evolution (DE) method.The aim of this study is to evaluate the performance of the new HMLSL algorithm on the testbed of 24 noiseless functions. The new algorithm is also compared against a simple MLSL and a traditional DE in order to identify the benefits of the applied improvements.The results confirm that the HMLSL outperforms the MLSL and DE methods. The new method has a larger probability of success and usually is faster especially in the final stage of the optimization than the other two algorithms.

show abstract

“…In Fig. 5 and 6 the results of LGO are compared with the results of GLOBAL [4,16], MCS [13,14] and NEWUOA [26,31].…”

Section: Comparing Lgo With Global Mcs and Newuoamentioning

confidence: 99%

Benchmarking the LGO Solver Suite within the COCO Test Environment

Hatwagner

Pintér²

2014

Acta Tech Jaur

View full text Add to dashboard Cite

Optimization problems often arise in the context of scientific-engineering research and practice. In many situations that require optimization, there is no need to develop new, highly customized software for a new problem, because there are readily usable optimization packages to choose from. Benchmarking and testing software environments can greatly assist the process of choosing an appropriate optimization tool. The benchmarking environments typically include a substantial collection of well-known and widely used test functions; they also offer a properly defined methodology to compare the solvers considered. One of these benchmarking environments is called COCO (an abbreviation that stands for "COmparing Continuous Optimizers"). COCO has been used at the annual BBOB (Black-Box Optimization Benchmarking) workshops. COCO assesses the capabilities of optimization solvers based on a given collection of test problems and evaluation criteria, under identical and fully reproducible circumstances. The goal of our present study is to benchmark the LGO (Lipschitz Global Optimizer) solver suite and to compare it to several other solvers using COCO.

show abstract

Black Box Optimization Benchmarking of the GLOBAL Method

Cited by 16 publications

References 23 publications

SOMS: SurrOgate MultiStart algorithm for use with nonlinear programming for global optimization

SOMS: SurrOgate MultiStart algorithm for use with nonlinear programming for global optimization

Benchmarking a hybrid multi level single linkagealgorithm on the bbob noiseless testbed

Benchmarking the LGO Solver Suite within the COCO Test Environment

Contact Info

Product

Resources

About