Deadlock Prevention Policy with Behavioral Optimality or Suboptimality Achieved by the Redundancy Identification of Constraints and the Rearrangement of Monitors

Table 3

Performance comparison of policies applied to the example in Figure 5.

Policy referred to	Number of states	Number of monitors	Number of arcs	Complete siphon	Complete state	Computational complexity
Piroddi et al. [45]	21581	13	82	✓	✗	Exponential
Ezpeleta et al. [12]	6287	18	106	✓	✗	Exponential
Huang et al. [39]	12656	16	88	✗	✗	NP-hard
Huang [47]	16425	7	34	✓	✗	Exponential
Li and Zhou [14]	6287	6	32	✓	✗	Exponential
Li et al. [32]	6331	5	27	✗	✗	NP-hard
Li and Zhou [15]	15999	6	29	✗	✗	NP-hard
Park and Reveliotis [16]	2480	7	38	✓	✗	Polynomial
Tricas et al. [17]	14850	8	40	✗	✗	NP-hard
Uzam and Zhou [46]	21562	19	112	✗	✓	Exponential
Xing and Hu [48]	15098	4	17	✓	✗	Exponential
The proposed policy	20444	9	45	✗	✗	NP-hard