Performance Analysis of Trial and Error Algorithms

Abstract

Model-free decentralized optimizations and learning are receiving increasing attention from theoretical and practical perspectives. In particular, two fully decentralized learning algorithms, namely Trial and Error (TEL) and Optimal Dynamical Learning (ODL), are very appealing for a broad class of games. In fact, ODL has the property to spend a high proportion of time in an optimum state that maximizes the sum of utility of all players. And the TEL has the property to spend a high proportion of time in an optimum state that maximizes the sum of utility of all players if there is a Pure Nash Equilibrium (PNE), otherwise, it spends a high proportion of time in an optimum state that maximizes a tradeoff between the sum of utility of all players and a predefined stability function. On the other hand, estimating the mean fraction of time spent in the optimum state (as well as the mean time duration to reach it) is challenging due to the high complexity and dimension of the inherent Markov Chains.In this paper, under some specific system model, an evaluation of the above performance metrics is provided by proposing an approximation of the considered Markov chains, which allows overcoming the problem of high dimensionality. A comparison between the two algorithms is then performed which allows a better understanding of their performances.

Performance Analysis of Brushless DC Motor Using Modified Queen Bee Evolution Based Genetic Algorithm Tuned PI Controller under Different Speed Conditions

“…Values of the controller gain parameter K P and K I are obtained by trial and error method [10] for various reference speeds. During the manual tuning process system is kept operational.…”

Speed Control of Brushless DC Motor Using Modified Genetic Algorithm Tuned Fuzzy Controller

“…Trial and error method is used to obtain values of gain parameter K P and K I [8] for various reference speeds. Manual controller tuning, results in the following gain constant values.…”

“…Two parent chromosomes are randomly picked from the population and the probability of new chromosome creation, from the parents is determined by crossover rate. Numerous crossover operators are explained in various literatures [10,11], but the weight based 7and equation 8ore likely to be achieved by GA's, as it works with population of points, on contrary to point by point approach of (7) (8) Genetic algorithm comprises three basic operators called reproduction operator, crossover operator and mutation operator. Initially GA works with randomly created group of solutions, known as population.…”

Decentralized Q-Learning with Constant Aspirations in Stochastic Games

“…Although the particular revision mechanism may vary, it is typically characterized by a "win stay, lose shift" condition similar to that of (2), whereby unsatisfied agents consider randomized experimental action revision and satisfied agents continue using their previous baseline action. 2 A number of Nash-seeking algorithms and multi-agent learning algorithms from this family, including [14], [18], [20] and [27], have been studied by first analyzing the convergence properties of a Markov chain {a t } ∞ t=1 and then using {a t } ∞ t=1 to approximate a sequence of learned strategy iterates { a t } ∞ t=1 .…”