Date: March 17, 2024

Topic: Path vs Policy Search

Recall

Single path vs the best action at every state

Notes

Path vs Policy Search

Path search (left) vs policy search (right)

• A path search results in a single path produced from a single entry point
  • We can only utilize it at a single location
• A policy search gives the best action to take at every single state
  • We can utilize it at every location, although it is more computationally expensive

<aside> 📌 SUMMARY: Single path vs the best action at every state

</aside>

Date: March 17, 2024

Topic: Dynamic Programming

Recall

Notes

Dynamic Programming

Technique used to simplify a complicated problem by breaking it down into simpler sub-problems

• Optimal substructure:
  • An overall optimal solution can be created from the optimal solutions of sub-problems
• Overlapping sub-problems:
  • A problem that can be broken down into smaller problems is said to have sub-problems
  • If the same sub-problem exists more than once, then this is known as overlapping sub-problems

<aside> 📌 SUMMARY:

</aside>