Adaptive parameterized improving hit-and-run for global optimization

Authors:
Wei Wang;Archis Ghate;Zelda B. Zabinsky
Affiliations:
Industrial and Systems Engineering, University of Washington, Seattle, WA, USA;Industrial and Systems Engineering, University of Washington, Seattle, WA, USA;Industrial and Systems Engineering, University of Washington, Seattle, WA, USA
Venue:
Optimization Methods & Software - GLOBAL OPTIMIZATION
Year:
2009

Citing 20
Cited 1

Generalized simulated annealing for function optimization

Technometrics
Minimizing multimodal functions of continuous variables with the “simulated annealing” algorithm: Corrigenda for this article is available here

ACM Transactions on Mathematical Software (TOMS)
Markov chains with rare transitions and simulated annealing

Mathematics of Operations Research
Computational experience with generalized simulated annealing over continuous variables

American Journal of Mathematical and Management Sciences
Stochastic minimization of Lipschitz functions

Recent advances in global optimization
Pure adaptive search in global optimization

Mathematical Programming: Series A and B
Hit-and-run algorithms for generating multivariate distributions

Mathematics of Operations Research
Simulated annealing algorithms for continuous global optimization: convergence conditions

Journal of Optimization Theory and Applications
Convergence of the simulated annealing algorithm for continuous global optimization

Journal of Optimization Theory and Applications
Accelerating the convergence of random search methods for discrete stochastic optimization

ACM Transactions on Modeling and Computer Simulation (TOMACS)
A class of convergent generalized hill climbing algorithms

Applied Mathematics and Computation
Direction Choice for Accelerated Convergence in Hit-And-Run Sampling

Operations Research
Simulation Optimization: Integrating Research and Practice

INFORMS Journal on Computing
A Numerical Evaluation of Several Stochastic Algorithms on Selected Continuous Global Optimization Test Problems

Journal of Global Optimization
Univariate Global Optimization with Multiextremal Non-Differentiable Constraints Without Penalty Functions

Computational Optimization and Applications
Relative Frequencies of Generalized Simulated Annealing

Mathematics of Operations Research
Simulated Annealing for Convex Optimization

Mathematics of Operations Research
An analytically derived cooling schedule for simulated annealing

Journal of Global Optimization
A Hit-and-Run approach for generating scale invariant Small World networks

Networks
Adaptive search with stochastic acceptance probabilities for global optimization

Operations Research Letters

Global optimization

Optimization Methods & Software - GLOBAL OPTIMIZATION

Quantified Score

Hi-index	0.00

Visualization

Abstract

We build on improving hit-and-run's (IHR) prior success as a Monte Carlo random search algorithm for global optimization by generalizing the algorithm's sampling distribution. Specifically, in place of the uniform step-size distribution in IHR, we employ a family of parameterized step-size distributions to sample candidate points. The IHR step-size distribution is a special instance within this family. This parameterization is motivated by recent results on efficient decentralized search in the so-called Small World problems. To improve the performance of the algorithm, we adaptively tune the parameter based on the success rate of obtaining improving points. We present analytical and numerical results on simple spherical programmes to illustrate the key ideas of the relationship between the parametrization and algorithm performance. These results are then extended to global optimization problems with Lipschitz continuous objective functions. Our preliminary numerical results demonstrate the potential benefit of considering parameterized versions of IHR.