Reinforcement Learning for Optimization

Cumulative Reward

Sum of expected future rewards that the agent seeks to maximize, often calculated with a discount factor to give less weight to distant rewards.

← Wstecz