mohab metwally 24d68733ed typo fixed há 3 anos atrás
..
README.md 584dc5d3ad edit readme há 3 anos atrás
main.py 24d68733ed typo fixed há 3 anos atrás
target.png 24d68733ed typo fixed há 3 anos atrás

README.md

target function approximation

excluding use of floats, and division, only +,-,* are allowed.

consequences

  • hard coded tunning.
  • public reward function.

comparison of original target to approximation

alt text

conclusion

as the derivative of deltas graph shows, starting for term 2, the derivatives is ~ 0, and it's the optimal number of terms in approximation accuracy that has the least number of terms.