Q-Managed: A new algorithm for a multiobjective reinforcement learning.
Thiago Henrique Freire de OliveiraLuiz Paulo de Souza MedeirosAdrião Duarte Dória NetoJorge Dantas de MeloPublished in: Expert Syst. Appl. (2021)
Keyphrases
- reinforcement learning
- dynamic programming
- detection algorithm
- multi objective
- experimental evaluation
- learning algorithm
- optimization algorithm
- np hard
- computational cost
- matching algorithm
- k means
- preprocessing
- segmentation algorithm
- optimal solution
- probabilistic model
- monte carlo
- objective function
- expectation maximization
- lower bound
- computational complexity
- recognition algorithm
- significant improvement
- convergence rate
- test problems
- state space
- simulated annealing
- worst case
- cost function
- evolutionary algorithm