The Applicability of Reinforcement Learning Methods in the Development of Industry 4.0 Applications

Table 2

-test table of principle captured classes by RL method types.

Principle captured	Markov decision process	Multiarmed bandit	Dynamic	Temporal difference	Value function approximation	Policy gradient	Multiagent	Edge computing
Prediction, forecasting, estimation, planning	1.96	−0.58	0.12	−0.32	3.08	0.31	−0.39	−4.17
Detection, recognition, prevention, avoidance, protection	−3.09	1.51	−1.11	−0.17	0.9988, 0.9782, 0.97850.39	0.9935, 0.8769, 0.87852.2	−1.38	1.65
Evaluation, assessment	−0.82	0.63	−0.08	−0.62	−1.21	−0.6	2.96	−0.27
Classification, clustering	−3.61	0.56	1.28	1.41	1.54	0.65	−2.88	1.05
Decision making	3.1	2.47	−0.84	1.73	0.95	0.02	−10.82	3.39
Allocation, assignment, resource management	−2.69	−1.95	0.21	−11.18	−4.74	2.08	10.54	7.74
Scheduling, queuing, planning	2.17	−1.33	−2.17	1.24	1.61	−0.04	−2.63	1.16
Control	2.67	−1.34	2.35	7.52	−1.72	−4.93	4.27	−10.82