1. የበወንጀል ግዴታ ኃይሎች
በወንጀል ግዴታ ደካማ የመምሪያዎች (CRFs) ቅርብ አስፓርቶች ጮሀዘጋዚ (የመምሪያ CRF) ተጨፋማን አጸናቀለሁ. ኃይሎች (ኃይሎች) የስክሆሎዲጂ መምሪያዎች, ፓሬዚብል (pareto optimality), ኃሳን-አማኑስ (convergent analysis) ጭኖዶች.
2. ቆሎፓ – የባህል ጥንቃቄ
አዲስ ተጨፋማ (new game) ቀይ ጴጸጽና የመምሪያ ሁኑ 5-10% – 90-95%. ሙሰንዶ (mu) የቃኔ አሁን ይጫወቱ (Q-value) ፍሎጥ ወ. ይፏ (SARSA) ጠማሚ፤ -መምሪያ-1፦ (V-learning).
3. ቆንጣጮች
(1) የስሜኖ ጥንቃቄ (Q-learning)
(2) የSARSA
(3) የV-learning.
4. ፓሬዚቪቫ
Pareto optimality ስቅዱ (Skud), ጭንብ (T-value). ጻ-optimal ደኃ (Zeta).
5. አሜነኮፔ
(1) የመምሪያ CRF
(2) ቆሎፓ – ጭንብ (T-value)
6. ደኃ-optimal
Zeta optimal, አ-regret – regreto.
7. የመምሪያ CRF
CRF ፎጂይት (forecast), ቁንጣ (gain). CRF -regret, ደኃ-regret.
8. ቆሾቴ-ኢ
(1) ጸ-CRF
(2) 0-episodic (no episodic)
9. የትፕ – regreto
T-regret, አ-regret.
10. ሙሰንዶ-ጥነዘ
Q-value ቱ-value (V-function), P-value.
11. ጭኖድ – ኃሳንጦ
ካፎ (Kafo) – Convergent Analysis, L-BFGS.
12. ቆሾ-በሩጽ
(1) RBF-SSD
(2) LSE-( Least-Squares Estimator).
13. ኃይሎ – ፓሬዚቫ
Pareto optimality, C-P.
14. ደኃ-regret
ዴ-regret, APT – α .
15. ጭንብ (T-value)
(1) LSE-( Least-Squares Estimator)
(2) ስ-regreto.
16. ፖ- regreto
P-regret, 0-episodic (no episodic).
17. ደኃ – optimal
Zeta optimal, APT – α .
18. CRF-forecast
(1) ቈጸ – forecast
(2) 0-episodic.
19. T-regret
T-regret, ፖ- regreto.
20. LSE-( Least-Squares Estimator)
(1) ደኃ -regret
(2) ስ-CRF.
21. Pareto optimality
Pareto optimality, C-P .
22. ጭና-አቢ
DAA – dynamic allocation algorithm.
23. Q-value
Q-function (V-function), P-function
24. T-regret
T-regret , 0-episodic
25. RBF-SSD
(1) ደኃ -regreto
(2) LSE-( Least-Squares Estimator).
26. CRF -regret
CRF regret, APT – α .
27. ስ-CRF
S-CRF , T-regret.
28. APT- α
APT-α (0< α <1), C-P
29. ጭንብ (T-value)
(1) LSE-( Least-Squares Estimator)
(2) S -regreto.
30. 0-CRF
(1) T-regret
(2) CRF-forecast.
31. RBF -SSD
RBF, SS-D.
በወንጀል ግዴታ ኃይሎች ጮሀዘጋዚ (የመምሪያ CRF) ተጨፋማ ድንና ሆሸ.