DocumentCode :
1242399
Title :
Learning the global maximum with parameterized learning automata
Author :
Thathachar, M. A L ; Phansalkar, V.V.
Author_Institution :
Dept. of Electr. Eng., Indian Inst. of Sci., Bangalore, India
Volume :
6
Issue :
2
fYear :
1995
fDate :
3/1/1995 12:00:00 AM
Firstpage :
398
Lastpage :
406
Abstract :
A feedforward network composed of units of teams of parameterized learning automata is considered as a model of a reinforcement learning system. The internal state vector of each learning automaton is updated using an algorithm consisting of a gradient-following term and a random perturbation term. It is shown that the algorithm weakly converges to a solution of the Langevin equation, implying that the algorithm globally maximizes an appropriate function. The algorithm is decentralized, and the units do not have any information exchange during updating. Simulation results on common payoff games and pattern recognition problems show that reasonable rates of convergence can be obtained
Keywords :
convergence; feedforward neural nets; game theory; learning (artificial intelligence); learning automata; optimisation; pattern recognition; simulation; Langevin equation; convergence rates; decentralized algorithm; feedforward network; global maximum; gradient-following term; internal state vector updating; parameterized learning automata; pattern recognition problems; payoff games; random perturbation term; reinforcement learning system; weakly converging algorithm; Convergence; Equations; Learning automata; Neural networks; Optimization methods; Pattern recognition; Random variables; Robustness; Simulated annealing; Tunneling;
fLanguage :
English
Journal_Title :
Neural Networks, IEEE Transactions on
Publisher :
ieee
ISSN :
1045-9227
Type :
jour
DOI :
10.1109/72.363475
Filename :
363475
Link To Document :
بازگشت