The NEWUOA software for unconstrained optimization without derivatives

By M. Powell. 2006

555 citations recorded by the source corpus.

Subfields: Computational Theory and Mathematics, Numerical Analysis

View Paper: https://www.semanticscholar.org/paper/d32a18f84491b7ba90cc8d736f63b08573b4cb6c

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Theory and Mathematics

In Numerical Analysis

Citations in the corpus

COCO: a platform for comparing continuous optimizers in a black-box setting (2016, methodology): “This call processes the data sets of seven (more-or-less representative) solvers benchmarked on the 24 unconstrained continuous test functions of the bbob test suite: • NEWUOA, the NEW Unconstrained Optimization Algorithm [59] with the recommended number of 2n + 1 interpolation points in the quadrat”
Derivative-free optimization: a review of algorithms and comparison of software implementations (2012, methodology): “Powell [113] proposed an algorithm that uses a quadratic model relying on fewer than (n +1)(n +2)/2 interpolation points.”
A New Method for Improving Functional-to-Structural MRI Alignment using Local Pearson Correlation (2008, methodology): “To carry out the minimization of C[E, S] over the parameter vector θ, our software incorporates Powell's derivative-free NEWUOA software (Powell, 2006), modified slightly to allow for range constraints on the solution parameters.”
Benchmarking Derivative-Free Optimization Algorithms (2009, methodology): “For example, a typical convergence test for trust region methods [1, 14, 16] requires that the trust region radius be smaller than a given tolerance.”
Aggregation Functions: A Guide for Practitioners (2007, methodology): “There are a number of derivative-free methods of non-smooth optimization [9, 37, 205, 206], and we particularly mention the Bundle methods [9, 139, 154].”
Recent advances in trust region algorithms (2015, background): “Powell [97–99] suggests which is used in a series of software [100–102].”
A View of Algorithms for Optimization without Derivatives 1 (2007, methodology): “This updating method is at the heart of my NEWUOA Fortran software (Powell, 2006), which has superseded UOBYQA.”
CUTEst: a Constrained and Unconstrained Testing Environment with safe threads for mathematical optimization (2013, unknown): “In addition to the still-current packages supported by CUTEr, CUTEst provides new interfaces to ALGENCAN [1], BOBYQA [19], Direct Search [7], filterSD [8], NEWUOA [18], NLPQLP [21], NOMAD [15], PENNLP [14], QL [20], SPG [2] and SQIC [10], as well as various new packages within GALAHAD [12].”
MAMPOSSt: Modelling Anisotropy and Mass Profiles of Observed Spherical Systems – I. Gaussian 3D velocities (2012, methodology): “To search for the best-fit solution, we run the MAMPOSSt algorithm in combination with the NEWUOA4 minimization routine of Powell (2006).”
CLASH-VLT: The mass, velocity-anisotropy, and pseudo-phase-space density profiles of the z = 0.44 galaxy cluster MACS J1206.2-0847 (2013, methodology): “The best-fits are obtained using the NEWUOA software for unconstrained optimization (Powell 2006).”
Bibliography on stable distributions, processes and related topics (2004, background): “New York: The Clarendon Press Oxford University Press. A Bayesian tutorial. Skorohod, V. (1957). Limit theorems for stochastic processes with independent increments.”
Reinforcement Learning in Continuous State and Action Spaces (2012, background): “Another approach is to construct an easier solvable (e.g., quadratic)model of the function that is to be optimized and then maximize this model analytically(see, e.g., Powell, 2002, 2006; Huyer and Neumaier, 2008).”
Unifying Optimization Algorithms to Aid Software System Users: optimx for R (2011, methodology): “Another development has been the interfacing of several quadratic approximation approaches of Powell (2002, 2006, 2008) by Katherine Mullen (aided by John Nash, Douglas Bates and Ravi Varadhan) in the form of the minqa package (Bates et al. 2011).”
The BOBYQA algorithm for bound constrained optimization without derivatives (2009, methodology): “The details of a shift are taken from NEWUOA (Powell, 2006).”