Ant colony optimization and stochastic gradient descent
Artificial Life
Intelligent systems: architectures and perspectives
Recent advances in intelligent paradigms and applications
A learning automata based algorithm for optimization of continuous complex functions
Information Sciences: an International Journal
A learning automata based algorithm for optimization of continuous complex functions
Information Sciences: an International Journal
IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews
Generalized learning automata for multi-agent reinforcement learning
AI Communications - European Workshop on Multi-Agent Systems (EUMAS) 2009
MLDM'05 Proceedings of the 4th international conference on Machine Learning and Data Mining in Pattern Recognition
Hi-index | 0.00 |
This paper analyzes the long-term behavior of the REINFORCE andrelated algorithms (Williams 1986, 1988, 1992) for generalizedlearning automata (Narendra and Thathachar 1989) for theassociative reinforcement learning problem (Barto and Anandan1985). The learning system considered here is a feedforwardconnectionist network of generalized learning automata units. Weshow that REINFORCE is a gradient ascent algorithm but can exhibitunbounded behavior. A modified version of this algorithm, based onconstrained optimization techniques, is suggested to overcome thisdisadvantage. The modified algorithm is shown to exhibit localoptimization properties. A global version of the algorithm, basedon constant temperature heat bath techniques, is also described andshown to converge to the global maximum. All algorithms areanalyzed using weak convergence techniques.