Comparison of the SPSA and simulated annealing algorithms for the constrained optimization of discrete non-separable functions

• Published in: 2003 Conference American Control Vol. 4; pp. 3260 - 3262 vol.4
• Main Authors: Whitney, J.E., Hill, S.D., Wairia, D., Bahari, F.
• Format: Conference Proceeding
• Language: English
• Published: IEEE 2003
• Subjects: Approximation algorithms; Constraint optimization; Iterative algorithms; Laboratories; Noise measurement; Particle measurements; Physics; Resource management; Simulated annealing; Stochastic processes
• ISBN: 9780780378964; 0780378962
• ISSN: 0743-1619
• DOI: 10.1109/ACC.2003.1244033

This paper considers a version of the simultaneous perturbation stochastic approximation (SPSA) algorithm and a simulated annealing (SAN) algorithm for optimizing non-separable functions over discrete sets under given constraints. The primary motivation for the application of these discrete optimization algorithms is to solve a class of resource allocation problems wherein the goal is to distribute a finite number of discrete resources to finitely many users in such a way as to optimize a specified objective function. The basic algorithms and their application to the optimal resource allocation problem are discussed and simulation results are presented which compare their performance.