SIR 실험 계획

[boyeon-kim]

• SIR 1) initial setting
  • S0 = 9,999,990
  • I0 = 10
  • beta = 0.000000036
  • gamma = 1/10
  • Constraint (number nu)
    • daily : 500,000
    • Total : 5,000,000
  • time : 300
  • Reward : - I - nu
  • penalty : if sum(nu) > total : penalty reward - 200,000
  • Reward scaling : reward/1e7

[boyeon-kim]

Result

[boyeon-kim]

2nd todo

• penalty term 작동하는지 확인

[boyeon-kim]

이거 close 해도 될 것 같습니다!