A General Self-Adaptive Relaxed-PPA Method for Convex Programming with Linear Constraints

Table 2

Performance of different iterative methods.

𝐴	TFOCS		ADMM		PPA-CM		SPDRPPAG1-I
𝑛	It./CPU	Err.	It./CPU	Err.	It./CPU	Err.	It./CPU	Err.
512	1681/4.16	7.5𝑒−10	492/0.67	5.0𝑒−10	1027/1.41	6.2𝑒−10	391/0.10	4.0𝑒−10
1024	406/1.68	3.7𝑒−10	274/0.53	3.0𝑒−10	848/0.76	2.3𝑒−10	214/0.28	3.0𝑒−10
2048	507/4.01	4.4𝑒−9	607/3.93	2.9𝑒−9	804/3.65	4.0𝑒−9	239/1.49	3.9𝑒−9
4096	933/19.50	2.6𝑒−9	1070/27.47	2.1𝑒−9	845/15.09	2.3𝑒−9	422/9.84	2.3𝑒−9
4200	461/9.91	2.1𝑒−9	916/25.48	1.8𝑒−9	868/16.27	2.1𝑒−9	391/9.44	1.5𝑒−9
4300	451/9.74	2.2𝑒−9	464/20.37	1.6𝑒−9	884/16.91	2.1𝑒−9	429/10.54	2.1𝑒−9
4600	505/12.37	1.36𝑒−9	2155/56.45	1.2𝑒−9	863/18.95	1.2𝑒−9	425/11.89	1.1𝑒−9
4700	801/20.15	1.3𝑒−11	1517/45.37	8.2𝑒−12	1102/34.27	1.2𝑒−11	425/14.20	1.1𝑒−11
5500	407/13.64	2.4𝑒−6	—/—	—	546/20.13	1.9𝑒−6	308/12.40	2.0𝑒−6
6500	1257/56.11	1.5𝑒−5	—/—	—	508/26.48	1.4𝑒−5	308/17.22	1.3𝑒−5
7500	801/81.42	1.1𝑒−12	—/—	—	1313/813.52	1.1𝑒−12	522/69.52	1.1𝑒−12
8500	842/107.18	2.3𝑒−7	—/—	—	724/100.36	2.2𝑒−7	542/83.07	2.2𝑒−7