Empirical analysis of the convergence of Double DQN in relation to reward sparsity.

Samuel Blad Martin Längkvist Franziska Klügl Amy Loutfi

Published in: ICMLA (2022)

Keyphrases

empirical analysis
theoretical analysis
empirical studies
convergence speed
reinforcement learning
convergence rate
sparse representation
long run
listed companies
neural network
high dimensional
long term
initial conditions
faster convergence
optimal solution