Multi-Objective reward generalization: Improving performance of Deep Reinforcement Learning for selected applications in stock and cryptocurrency trading.
Federico CornalbaConstantin DisselkampDavide ScassolaChristopher HelfPublished in: CoRR (2022)
Keyphrases
- reinforcement learning
- multi objective
- trading systems
- stock trading
- stock data
- multi objective optimization
- evolutionary algorithm
- optimization algorithm
- function approximation
- state space
- reinforcement learning algorithms
- eligibility traces
- particle swarm optimization
- reward function
- markov decision processes
- genetic algorithm
- temporal difference
- partially observable environments
- nsga ii
- stock market
- stock price
- learning algorithm
- multi objective optimization problems
- multi objective evolutionary
- optimal policy
- electronic commerce
- financial markets
- learning agent
- dynamic programming
- conflicting objectives
- multiple objectives
- optimization problems
- financial information
- multi objective evolutionary algorithms
- average reward
- learning process
- multi agent
- agent receives
- policy evaluation
- state action
- bi objective
- model free
- pareto optimal
- differential evolution