selected-solutions.md

Selected Reference Solution For Chapter 2

Exercise 2.1

Exercise 2.2

Definitely occured: time step 2 and 5; might occur:time step 1, 3 and 4;

Exercise 2.3

epsilon = 0.01 will perform better in long run.

Exercise 2.4

Exercise 2.7

Exercise 2.8

Case A or B unknown, best expectation = 0.5; behave randomly will achieve Case A or B known, best expectation = 0.55; For case A, choose action 2; whereas case B, choose action 1;

---

Experiments Results For Chapter 2

Bandit Problem

Average Rewards For Different methods

Optimal Actions For Different methods