Title :
Stochastic approximation for discrete optimization of noisy loss measurements
Author_Institution :
Dept. of Appl. Math. & Stat., Johns Hopkins Univ., Baltimore, MD, USA
Abstract :
We consider the stochastic optimization of a noisy convex loss function defined on p-dimensional grid of points in Euclidean space. We introduce the middle point discrete simultaneous perturbation stochastic approximation (DSPSA) algorithm to this discrete space. Consistent with other stochastic approximation methods, this method formally accommodates noisy measurements of the loss function.
Keywords :
approximation theory; optimisation; stochastic processes; Euclidean space; discrete optimization; discrete space; middle point discrete simultaneous perturbation stochastic approximation algorithm; noisy convex loss function; noisy measurement; p-dimensional point grid; stochastic optimization; Approximation methods; Convergence; Convex functions; Loss measurement; Noise measurement; Optimization; Search methods; SPSA; Stochastic optimization; discrete optimization; noisy data; recursive estimation;
Conference_Titel :
Information Sciences and Systems (CISS), 2011 45th Annual Conference on
Conference_Location :
Baltimore, MD
Print_ISBN :
978-1-4244-9846-8
Electronic_ISBN :
978-1-4244-9847-5
DOI :
10.1109/CISS.2011.5766217
