An adaptive search algorithm for numerical optimization
This paper takes a new look at the popular simplex method of \textit{A. J. Nelder} and \textit{R. Mead} [(*) A simplex method for function minimization, Comput. J. 308 (1964)], with which we compute minima of functions of several variables without computing any first or second derivatives. A section of the paper carefully and clearly reviews the simplex method (*); it then presents two improvements to the basic algorithm. Finally, the results of numerical experiments appear. The algorithm (*) involves reflection, expansion, and contraction steps. The authors replace the expansion and reflection by a special line search step. The line search itself is interesting, since it involves using Fibonacci ratios both to first expand the original interval to bracket the minimum and then to contract the bracketing interval. The line search makes the algorithm (*) more efficient and less sensitive to scaling. Viewed another way, it relieves the user of some of the burden of ``tuning the algorithm. The second proposed modification is use of a stochastic algorithm to determine a starting simplex for the algorithm (*), in order to converge to a global minimum instead of just to local one. The numerical tests include Rosenbrock's functions and variants, and a special multimodal function. Besides being clearly written, the paper is well illustrated.
- An efficient interfacing of golden-section and quadratic searches
- Efficient optimization of certain functionals in a hypercube
- Numerical experiments with the one-dimensional non-linear simplex search
- Gilding the lily: A variant of the Nelder-Mead algorithm based on Golden-section search
- Accelerating optimization by tracing valley
- Stochastic minimization with adaptive memory
- Derivative free optimization in higher dimension
- scientific article; zbMATH DE number 1728444 (Why is no real title available?)
- A revised simplex search procedure for stochastic simulation response surface optimization
- scientific article; zbMATH DE number 5932201 (Why is no real title available?)
- A modified simplex method
- A global optimization method based on the reduced simplicial statistical model
- Efficient Implementation of the Nelder-Mead Search Algorithm
- scientific article; zbMATH DE number 4179150 (Why is no real title available?)
- scientific article; zbMATH DE number 3848513 (Why is no real title available?)
- Convergent simplex searches and the ‘gloveless DataGlove’
- The use of hypothetical points in numerical optimization
- scientific article; zbMATH DE number 3952512 (Why is no real title available?)
- Numerical methods of minimization of nonlinear functions on disjoint simplices
- scientific article; zbMATH DE number 1187210 (Why is no real title available?)
- scientific article; zbMATH DE number 14314 (Why is no real title available?)
- scientific article; zbMATH DE number 1222768 (Why is no real title available?)
- scientific article; zbMATH DE number 1301900 (Why is no real title available?)
- scientific article; zbMATH DE number 515396 (Why is no real title available?)
- Nelder-Mead Simplex Modifications for Simulation Optimization
- Some modifications of low-dimensional simplex evolution and their convergence
- Fortified-Descent Simplicial Search Method: A General Approach
- Multistart global optimization algorithm based on simplex gradient
- scientific article; zbMATH DE number 973930 (Why is no real title available?)
- A recursive direct search algorithm for a minimization of multidimensional function
- Simple Metaheuristics Using the Simplex Algorithm for Non-linear Programming
- One-dimensional simplex search
- A combined topographical search strategy with ellipsometric application
- Searching minima of an N-dimensional surface: A robust valley following method
This page was built for publication: An adaptive search algorithm for numerical optimization
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q1100854)
