BoxPusher (Sokoban) solved: Reinforcement Learning vs. BFS

(Problem definition from LeetCode challenge 1263)

RL solution in "RLutils.h"

• Simulated environment: RL::RLField
• Policy class: RL::EpsGreedyPolicy
• Q-learning algo: RL::RL_solve

BFS solution in "BFSutils.h"

Demo:

• RL Solution using a (-1) reward for each step and (+100) for reaching the target
  
  
• RL Solution using a (-1) reward for each step, (-5) for each box move and (+100) for reaching the target