Table 2.

Performance of PIRN_CH on 100 Simulated Datasets Per Setting

	n = 30			n = 40*			n = 50*
	K = 3	K = 4	K = 5	K = 3	K = 4	K = 5	K = 3	K = 4	K = 5
= RH	98	93	77	97	90	83	98	89	82
SIT = RH	97	92	78	92	73	55	96	75	58
Gap(RH)	0.02	0.08	0.25	0.03	0.11	0.18	0.02	0.10	0.18
<SIT	1	3	3	5	22	37	2	16	34
=SIT	99	97	96	95	78	63	98	84	66
>SIT	0	0	1	0	0	0	0	0	0
Gap(SIT)	0.01	0.03	0.02	0.06	0.25	0.54	0.02	0.17	0.39
Time	850.6	3,321.3	6,453.6	2,942.7	5,299.8	16,384.3	2073.7	8,204.7	13,846.64

= RH (resp. SIT = RH): the number of datasets PIRN_CH (resp. SIT bound) gives the same results as the RH lower bound (and thus optimal networks are found). ^*: coarse mode of the SIT bound is used for n = 40 and 50. Gap(RH): average gap between PIRN_CH results and the RH bound. < SIT (the other two are straightforward): the number of datasets PIRN_CH gives the smaller hybridization number as given by the SIT upper bound. Gap(SIT): average gap between PIRN_CH results and the SIT bound. Gap of two values a and b is defined as a − b. Time: the time of PIRN_CH in seconds.