Juliane Mueller

Postdoctoral Researcher, Computational Research Division


Contact Information

Juliane Mueller
MS 50A-3111
Lawrence Berkeley National Lab
1 Cyclotron Rd.
Berkeley, CA 94720

510-495-2872 (ofc)
510-486-6900 (fax)

JulianeMueller@lbl.gov


Affiliation and Research Interests

I am an Alvarez Postdoctoral Fellow in the Computational Research Division of the Computing Sciences Directorate at the Lawrence Berkeley National Laboratory, and affiliated with the Center for Computational Sciences and Engineering (CCSE).

My research is in numerical optimization, especially the development of efficient derivative-free global optimization algorithms for finding (near) optimal solutions for computationally expensive black-box problems. I am using surrogate/response surface models. I work on continuous, discrete, and mixed-integer optimization problems that may have computationally expensive black-box constraints. In the course of my research I addressed applications arising in vehicle routing, structural optimization, renewable energy (hydropower, kite energy), watershed management, optimal reliability design, and global climate models.


For further information on optimization research conducted in our group, please see our Optimization Research Website.

Publications

J. Mueller, R. Paudel, C.A. Shoemaker, J. Woodbury, Y. Wang, N. Mahowald, "CH4 parameter estimation in CLM4.5bgc using surrogate global optimization", Geoscientific Model Development Discussions, 8, 141-207, 2015. [link]

J. Mueller, T. Krityakierne, C.A. Shoemaker, " SO-MODS: Optimization for high dimensional computationally expensive multi-modal functions with surrogate search", 2014 IEEE Congress on Evolutionary Computation (CEC), July 6-11, 2014, Beijing, China, 1092-1099, 2014. [link]

J. Mueller, C.A. Shoemaker, "Influence of Ensemble Surrogate Models and Sampling Strategy on the Solution Quality of Algorithms for Computationally Expensive Black-box Global Optimization Problems", Journal of Global Optimization, 60, 123-144, 2014. [link]

J. Mueller, C.A. Shoemaker, R. Piche, "SO-MI: A Surrogate Model Algorithm for Computationally Expensive Nonlinear Mixed-Integer Black-Box Global Optimization Problems", Computers and Operations Research, 40, 1383-1400, 2013. [link]

J. Mueller, J. Kanniainen, R. Piche, "Calibration of GARCH Models Using Concurrent Accelerated Random Search", Applied Mathematics and Computation, 221, 522-534, 2013. [link]

J. Mueller, C.A. Shoemaker, R. Piche, "SO-I: a surrogate model algorithm for expensive nonlinear integer programming problems including global optimization applications", Journal of Global Optimization, 59, 865-889, 2014. [link]

J. Mueller, R. Piche, "Mixture surrogate models based on Dempster-Shafer theory for global optimization problems", Journal of Global Optimization, 51, 79-104, 2010. [link]

J. Mueller, "Approximative solutions to the Bicriterion Vehicle Routing Problem with Time Windows", European Journal of Operational Research, 202, 223-231, 2010. [link]



Scientific Posters

J. Mueller, R. Paudel, C.A. Shoemaker, J. Woodbury, Y. Wang, N. Mahowald, "Parameter estimation in CLM4.5 using surrogate model based global optimization", AGU 2014 Fall Meeting, San Francisco, CA, USA. [pdf]

J. Mueller, "Mixture surrogate models based on Dempster-Shafer theory for global optimization", SULATIS (Finnish Society for Computational Sciences), Computational Sciences Day, September 2010, Kuopio, Finland. [pdf]



Codes for Surrogate Model Based Optimization

MISO Mixed Integer Surrogate Optimization framework: A MATLAB implementation of a surrogate model algorithm for computationally expensive mixed-integer black-box optimization problems with box constraints. Allows to choose from different radial basis function types, sampling strategies, and initial experimental design options.
MISO (MATLAB only) code click to download the zip archive
MISO Matlab manual
Journal article: J. Mueller. MISO: Mixed-Integer Surrogate Optimization framework, Optimization and Engineering, 17(1): 177-203, 2016

MATSuMoTo MATLAB Surrogate Model Toolbox: Surrogate model toolbox for box-constrained global optimization problems (continuous, pure integer, mixed-integer). Contains various surrogate model mixtures, initial experimental design strategies, and sampling strategies
MATSuMoTo (MATLAB only) code
MATSuMoTo Matlab manual
J. Mueller. MATSuMoTo: The MATLAB Surrogate Model Toolbox for Computationally Expensive Black-Box Global Optimization Problems, arXiv:1404.4261, April, 2014

Stochastic RBF codes: surrogate model optimization algorithm applicable for computationally expensive, black-box global optimization problems. MATLAB version requires MATLAB 2010b or newer. For doing several evaluations in each iteration, MATLAB Parallel Computing Toolbox is required. Python version requires Python 2.7.
Stochastic RBF Python code
Stochastic RBF Python manual
Stochastic RBF MATLAB code
Stochastic RBF MATLAB manual

DYCORS codes: surrogate model optimization algorithm applicable for computationally expensive, black-box global optimization problems with large dimensions (>30). MATLAB version requires MATLAB 2010b or newer. For doing several evaluations in each iteration, MATLAB Parallel Computing Toolbox is required. Python version requires Python 2.7.
DYCORS Python code
DYCORS Python manual
DYCORS MATLAB code
DYCORS MATLAB manual