AISpace2 Main Tools News Downloads Prototype Tools Customizable Applets Practice Exercises Help About AIspace

What is one potential problem with a Q-learning agent that always chooses the action which maximizes the Q-value? The agent might be stuck performing non-optimal actions. Unless it has already found the optimal policy, exploring actions which do not have the highest Q-value might allow it to find a better policy.

Main Tools: Graph Searching | Consistency for CSP | SLS for CSP | Deduction | Belief and Decision Networks | Decision Trees | Neural Networks | STRIPS to CSP