A Nonmonotone Trust Region Algorithm Based on the Average of the Successive Penalty Function Values for Nonlinear Optimization

Table 1

Test results for our method and the methods in [15, 18].

No.	Our method		The method in [18]		The method in [15]
	𝐼𝑡/𝐼𝑔	Time	𝐼𝑡/𝐼𝑔	Time	𝐼𝑡/𝐼𝑔	Time
H28	11/13	0.3438	13/24	0.2652	13/24	0.1404
H39	59/61	1.4688	24/37	0.3432	64/126	0.2625
H42	45/73	0.8281	133/195	0.2652	fail
H47	17/21	0.5625	63/121	0.5460	60/118	0.2964
H48	7/10	0.3125	14/26	0.2340	14/26	0.7488
H49	100/197	2.3750	118/234	0.4524	118/234	0.3144
H50	23/27	0.7344	63/124	0.9360	63/124	0.4212
H51	143/223	2.6094	57/88	0.6084	57/88	1.0764
H52	426/658	7.4844	50/100	0.8112	149/188	1.9812
H63	18/20	0.5469	15/27	0.5928	15/27	0.1404
H77	11/15	0.4063	25/48	1.2324	109/132	3.4788