论文信息 - Rate of convergence analysis of discrete simultaneous perturbation stochastic approximation algorithm

Rate of convergence analysis of discrete simultaneous perturbation stochastic approximation algorithm

A middle point discrete version of simultaneous perturbation stochastic approximation (DSPSA) algorithm was previously introduced to solve the discrete stochastic optimization problem. We consider the rate of convergence of DSPSA in this paper. This rate will allow for objective comparisons with other discrete stochastic optimization methods.

Qi Wang | James C. Spall

[1] L. Gerencsér,et al. Stochastic approximation on discrete sets using simultaneous difference approximations , 2004, Proceedings of the 2004 American Control Conference.

[2] James C. Spall,et al. Introduction to stochastic search and optimization - estimation, simulation, and control , 2003, Wiley-Interscience series in discrete mathematics and optimization.

[3] S. Andradóttir. A method for discrete stochastic optimization , 1995 .

[4] J. Spall. Multivariate stochastic approximation using a simultaneous perturbation gradient approximation , 1992 .

[5] S. Andradóttir,et al. A Simulated Annealing Algorithm with Constant Temperature for Discrete Stochastic Optimization , 1999 .

[6] Yu-Chi Ho,et al. Stochastic Comparison Algorithm for Discrete Optimization with Estimation , 1999, SIAM J. Optim..

[7] Barry L. Nelson,et al. Discrete Optimization via Simulation Using COMPASS , 2006, Oper. Res..

[8] Pavel Popela,et al. Integer simulation based optimization by local search , 2010, ICCS.

[9] D. Yan,et al. Stochastic discrete optimization , 1992 .

[10] Discrete simultaneous perturbation stochastic approximation on loss function with noisy measurements , 2011, Proceedings of the 2011 American Control Conference.