Showing papers by "Paul Bourgine published in 1997"
Proceedings Article•
01 Dec 1997TL;DR: A RL algorithm is proposed based on a Finite-Difference method and proved to convergence to the optimal solution of the Hamilton-Jacobi-Bellman equation.
Abstract: This paper is concerned with the problem of Reinforcement Learning (RL) for continuous state space and time stochastic control problems. We state the Hamilton-Jacobi-Bellman equation satisfied by the value function and use a Finite-Difference method for designing a convergent approximation scheme. Then we propose a RL algorithm based on this scheme and prove its convergence to the optimal solution.
37 citations
01 Jan 1997
TL;DR: In this paper, a cognitive microeconomy is proposed to study the bounded rationality of specular and autoteaching agents in changing their hedonistic strategies in a society of such agents, where the compromise between exploration of new strategies and exploitation of better known old strategies can be managed by agents despite their bounded cognitive capacities and their limited experience capacities.
Abstract: Because humanity enters the era of knowledge, it becomes important to understand the role of immaterial investments in knowledge. That explains the importance of an macroeconomy of knowledge. As a first step in this direction, this paper proposes a cognitive microeconomy in human society. Humans are considered as specular, autoteaching and hedonistic agents. The cognitive microeconomy has for object to study the bounded rationality of specular and autoteaching agents in changing their hedonistic strategies in a society of such agents. One main question is how the compromise between exploration of new strategies and exploitation of better known old strategies can be managed by agents despite their bounded cognitive capacities and their limited experience capacities.