测量一组5层网络的迭代次数
如图左边5层网络很显然可以看作是右边的3层网络两个组合而成的,所以左边的5层网络的迭代次数和右边的3层网络的迭代次数有没有什么关系?
5层 |
3层 |
2*10*2*10*2 |
2*10*2 |
3*10*3*10*3 |
3*10*3 |
4*10*4*10*4 |
4*10*4 |
本文通过改变节点数制作了3对网络,比较两种网络迭代次数之间的关系。
制作一个5层的网络
这个网络的结构是3*10*3*10*3,输入3个值,输出3个值。在这个实验中让 输入固定都是0.1,输出固定都是(1,0,0)。将这个网络简写成
(0.1)-3*10*3*10*3-(3*k),k∈{0,1}
这个网络的收敛标准是
if (Math.abs(f4[0]-y[0])< δ && Math.abs(f4[1]-y[1])< δ && Math.abs(f4[2]-y[2])< δ )
具体进样顺序 |
|||
δ=0.5 |
迭代次数 |
||
0.1 |
1 |
判断是否达到收敛 |
|
梯度下降 |
|||
0.1 |
2 |
判断是否达到收敛 |
|
梯度下降 |
|||
…… |
|||
每当网路达到收敛标准记录迭代次数 |
|||
将这一过程重复199次 |
|||
δ=0.4 |
|||
…… |
|||
δ=1e-7 |
因为对应每个收敛标准δ都有一个特征的迭代次数n与之对应因此可以用迭代次数曲线n(δ)来评价网络性能。
本文尝试了δ从1e-7到0.5的共35个值。收敛时记录迭代次数n-3*10*3*10*3.
将这个过程重复199次取平均值为n。共收敛了35*199次。
本文还制作了两外两个网络
(0.1)-2*10*2*10*2-(2*k),k∈{0,1}
(0.1)-4*10*4*10*4-(4*k),k∈{0,1}
用同样的办法可以得到n-2*10*2*10*2和n-4*10*4*10*4.
得到表格
2*10*2*10*2 |
3*10*3*10*3 |
4*10*4*10*4 |
|||
δ |
迭代次数n |
迭代次数n |
迭代次数n |
d2/d4 |
d3/d4 |
0.5 |
1.934673367 |
1.959798995 |
3 |
0.644891 |
0.653266 |
0.4 |
5.371859296 |
5.341708543 |
7 |
0.767408 |
0.763101 |
0.3 |
10.27135678 |
10.18090452 |
11 |
0.93376 |
0.925537 |
0.2 |
18.66331658 |
18.52763819 |
20 |
0.933166 |
0.926382 |
0.1 |
40.10050251 |
38.36683417 |
39 |
1.028218 |
0.983765 |
0.01 |
325.6130653 |
268.5025126 |
237 |
1.373895 |
1.132922 |
0.001 |
2280.115578 |
1684.38191 |
1406 |
1.621704 |
1.197996 |
1.00E-04 |
15835.47739 |
12056.8191 |
10609 |
1.492646 |
1.136471 |
9.00E-05 |
17342.0603 |
13253.43719 |
11717 |
1.480077 |
1.131129 |
8.00E-05 |
19208.1206 |
14739.47739 |
13077 |
1.468848 |
1.12713 |
7.00E-05 |
21562.95477 |
16633.47236 |
14844 |
1.452638 |
1.120552 |
6.00E-05 |
24676.82412 |
19138.08543 |
17197 |
1.434949 |
1.112873 |
5.00E-05 |
28974.85427 |
22613.09045 |
20480 |
1.414788 |
1.104155 |
4.00E-05 |
35289.97487 |
27783.49749 |
25306 |
1.39453 |
1.097902 |
3.00E-05 |
45565.71357 |
36295.21106 |
33318 |
1.367601 |
1.089357 |
2.00E-05 |
65556.1005 |
53122.43719 |
49468 |
1.325222 |
1.073875 |
1.00E-05 |
123256.9196 |
102725.196 |
97258 |
1.267319 |
1.056213 |
9.00E-06 |
135849.9447 |
113718.9146 |
108216 |
1.255359 |
1.050851 |
8.00E-06 |
151453.6382 |
127307.3518 |
121518 |
1.246347 |
1.047642 |
7.00E-06 |
171409.3065 |
144868.4121 |
139247 |
1.230973 |
1.04037 |
6.00E-06 |
197814.8894 |
168071.0653 |
161784 |
1.22271 |
1.038861 |
5.00E-06 |
234464.0503 |
200668.6884 |
193666 |
1.210662 |
1.036159 |
4.00E-06 |
288965.3467 |
249317.9447 |
241691 |
1.195598 |
1.031557 |
3.00E-06 |
378816.8543 |
330180.2663 |
322513 |
1.174579 |
1.023774 |
2.00E-06 |
556550.6985 |
491437.4623 |
482398 |
1.153717 |
1.018739 |
1.00E-06 |
1080983.186 |
974155.5427 |
964074 |
1.121266 |
1.010457 |
9.00E-07 |
1208133.623 |
1080634.106 |
1072751 |
1.126201 |
1.007348 |
8.00E-07 |
1353904.367 |
1215191.643 |
1206559 |
1.12212 |
1.007155 |
7.00E-07 |
1539773.07 |
1387027.477 |
1381749 |
1.114365 |
1.00382 |
6.00E-07 |
1786506.508 |
1616167.709 |
1606431 |
1.112097 |
1.006061 |
5.00E-07 |
2133555.362 |
1937818.854 |
1931582 |
1.104564 |
1.003229 |
4.00E-07 |
2646545.613 |
2418741.794 |
2414921 |
1.095914 |
1.001582 |
3.00E-07 |
3500985.513 |
3222108.256 |
3226161 |
1.085186 |
0.998744 |
2.00E-07 |
5199169.422 |
4828569.704 |
4839273 |
1.07437 |
0.997788 |
1.00E-07 |
1.03E+07 |
9653872.432 |
9700878 |
1.057468 |
0.995155 |
当δ∈[4e-7,0.01]
d-2*10*2*10*2> d-3*10*3*10*3> d-4*10*4*10*4
但是当δ<4e-7时d-3*10*3*10*3< d-4*10*4*10*4
再制作一个三层的网络
比照5层网络将这个网络写成
(0.1)-3*10*3-(3*k),k∈{0,1}
用同样的办法可以得到迭代次数n-3*10*3
同样制作了另外两个3层的网络
(0.1)-2*10*2-(2*k),k∈{0,1}
(0.1)-4*10*4-(4*k),k∈{0,1}
可以得到迭代次数n-2*10*2和n-4*10*4
得到表格
2*10*2 |
3*10*3 |
4*10*4 |
|
|
|
δ |
迭代次数n |
迭代次数n |
迭代次数n |
d2/d4 |
d3/d4 |
0.5 |
1.834170854 |
2.150753769 |
4 |
0.458543 |
0.537688 |
0.4 |
5.467336683 |
5.668341709 |
7 |
0.781048 |
0.809763 |
0.3 |
10.3718593 |
10.52763819 |
12 |
0.864322 |
0.877303 |
0.2 |
18.98492462 |
19.09547739 |
21 |
0.904044 |
0.909308 |
0.1 |
41.81407035 |
41.92964824 |
43 |
0.97242 |
0.975108 |
0.01 |
405.9145729 |
400.9246231 |
397 |
1.022455 |
1.009886 |
0.001 |
3909.160804 |
3785.150754 |
3660 |
1.068077 |
1.034194 |
1.00E-04 |
38025.66332 |
35887.74372 |
33662 |
1.129632 |
1.06612 |
9.00E-05 |
42192.34171 |
39777.31156 |
37338 |
1.130011 |
1.065331 |
8.00E-05 |
47395.70352 |
44621.80905 |
41802 |
1.133814 |
1.067456 |
7.00E-05 |
54052.65327 |
50836.96482 |
47443 |
1.139318 |
1.071538 |
6.00E-05 |
62950.38191 |
59079.18593 |
55259 |
1.139188 |
1.069132 |
5.00E-05 |
75340.25628 |
70575.01005 |
65584 |
1.14876 |
1.076101 |
4.00E-05 |
93917.48744 |
87725.11558 |
81242 |
1.156021 |
1.0798 |
3.00E-05 |
124764.8342 |
116155.4975 |
107253 |
1.163276 |
1.083005 |
2.00E-05 |
186176.5779 |
172366.9598 |
158480 |
1.174764 |
1.087626 |
1.00E-05 |
368710.8643 |
338333.7889 |
308464 |
1.195312 |
1.096834 |
9.00E-06 |
408984.0804 |
375023.1005 |
341247 |
1.198499 |
1.098978 |
8.00E-06 |
459388.5276 |
420561.8643 |
382394 |
1.201349 |
1.099813 |
7.00E-06 |
524038.9497 |
478899.1457 |
435268 |
1.203945 |
1.10024 |
6.00E-06 |
609813.0704 |
556314.598 |
504206 |
1.209452 |
1.103348 |
5.00E-06 |
729919.3065 |
664199.0553 |
600920 |
1.21467 |
1.105304 |
4.00E-06 |
909130.6181 |
825530.8241 |
744853 |
1.22055 |
1.108314 |
3.00E-06 |
1207124.558 |
1091742.894 |
980381 |
1.231281 |
1.11359 |
2.00E-06 |
1799195.739 |
1619069.065 |
1449669 |
1.241108 |
1.116854 |
1.00E-06 |
3559300.412 |
3175825.116 |
2818734 |
1.26273 |
1.126685 |
9.00E-07 |
3947484.774 |
3517441.141 |
3123123 |
1.263954 |
1.126258 |
8.00E-07 |
4433661.618 |
3943425.156 |
3493458 |
1.269133 |
1.128803 |
7.00E-07 |
5056311.834 |
4490016.492 |
3972648 |
1.272781 |
1.130233 |
6.00E-07 |
5882920.477 |
5216202.241 |
4605197 |
1.277453 |
1.132677 |
5.00E-07 |
7040336.97 |
6225307.176 |
5492530 |
1.281802 |
1.133413 |
4.00E-07 |
8766715.804 |
7731842.281 |
6800941 |
1.289045 |
1.136878 |
3.00E-07 |
1.16E+07 |
1.02E+07 |
8973350 |
1.296219 |
1.139355 |
2.00E-07 |
1.73E+07 |
1.52E+07 |
1.32E+07 |
1.308507 |
1.144752 |
1.00E-07 |
3.42E+07 |
2.97E+07 |
2.58E+07 |
1.327545 |
1.151904 |
这个结论是很清晰的
d-2*10*2>d-3*10*3>d-4*10*4.
比较两个网络
(0.1)-3*10*3*10*3-(3*k),k∈{0,1}
(0.1)-3*10*3-(3*k),k∈{0,1}
很显然这个5层的网络可以理解成是两个3层的网络组合成的,所以迭代次数n-3*10*3*10*3和n-3*10*3之间有什么关系?
δ |
d-2*10*2/d-2*10*2*10*2 |
d-3*10*3/d-3*10*3*10*3 |
d-4*10*4/d-4*10*4*10*4 |
0.5 |
0.948051948 |
1.097435897 |
1.333333333 |
0.4 |
1.01777362 |
1.061147695 |
1 |
0.3 |
1.009784736 |
1.034057256 |
1.090909091 |
0.2 |
1.017232095 |
1.030648223 |
1.05 |
0.1 |
1.04273183 |
1.092861821 |
1.102564103 |
0.01 |
1.246616356 |
1.493187603 |
1.675105485 |
0.001 |
1.714457303 |
2.247204587 |
2.603129445 |
1.00E-04 |
2.401295672 |
2.976551563 |
3.172966349 |
9.00E-05 |
2.432948622 |
3.001282686 |
3.186651873 |
8.00E-05 |
2.467482608 |
3.027367109 |
3.196604726 |
7.00E-05 |
2.506736847 |
3.056305005 |
3.196106171 |
6.00E-05 |
2.55099204 |
3.086995622 |
3.213293016 |
5.00E-05 |
2.600194485 |
3.12098031 |
3.20234375 |
4.00E-05 |
2.661307858 |
3.157454011 |
3.210384889 |
3.00E-05 |
2.738129712 |
3.200298169 |
3.219070773 |
2.00E-05 |
2.839958089 |
3.244711066 |
3.203687232 |
1.00E-05 |
2.991400933 |
3.293581343 |
3.171605421 |
9.00E-06 |
3.010557577 |
3.297807598 |
3.153387669 |
8.00E-06 |
3.033195723 |
3.303515928 |
3.146809526 |
7.00E-06 |
3.057237441 |
3.30575271 |
3.125869857 |
6.00E-06 |
3.082746056 |
3.309996262 |
3.1165381 |
5.00E-06 |
3.113139544 |
3.309928721 |
3.102867824 |
4.00E-06 |
3.146157933 |
3.311156865 |
3.081840035 |
3.00E-06 |
3.186565075 |
3.306505584 |
3.03981855 |
2.00E-06 |
3.232761622 |
3.294557679 |
3.005130618 |
1.00E-06 |
3.292651041 |
3.260080117 |
2.923773486 |
9.00E-07 |
3.267423982 |
3.254978834 |
2.911321453 |
8.00E-07 |
3.274722888 |
3.245105558 |
2.895389285 |
7.00E-07 |
3.283803264 |
3.237150356 |
2.875086575 |
6.00E-07 |
3.292974558 |
3.22751297 |
2.866725679 |
5.00E-07 |
3.299814524 |
3.212533082 |
2.843539648 |
4.00E-07 |
3.312512643 |
3.196638145 |
2.816216762 |
3.00E-07 |
3.322328032 |
3.173025172 |
2.781432793 |
2.00E-07 |
3.332789289 |
3.139484946 |
2.736437271 |
1.00E-07 |
3.33782706 |
3.077567328 |
2.658776041 |
1.00E-07 |
d-2*10*2 |
> |
d-2*10*2*10*2 |
3.33 |
1.00E-07 |
d-3*10*3 |
> |
d-3*10*3*10*3 |
3.07 |
1.00E-07 |
d-4*10*4 |
> |
d-4*10*4*10*4 |
2.65 |
一个大致的结论是在绝大多数区间上三层网路的迭代次数都要大于对应5层网络的迭代次数。比例并不固定但大于2倍。
学习率 0.1 |
权重初始化方式 |
Random rand1 =new Random(); |
int ti1=rand1.nextInt(98)+1; |
int xx=1; |
if(ti1%2==0) |
{ xx=-1;} |
tw[a][b]=xx*((double)ti1/1000);
|
d0.1-2-10-2-10-2
f2[0] f2[1] 迭代次数n 平均准确率p-ave δ 耗时ms/次 耗时ms/199次 耗时 min/199
0.528641 0.470845 1.934673 0 0.5 0.472362 94 0.001567
0.620223 0.38049 5.371859 0 0.4 0.155779 31 0.000517
0.713105 0.28614 10.27136 0 0.3 0.236181 47 0.000783
0.806761 0.192324 18.66332 0 0.2 0.155779 47 0.000783
0.902126 0.098016 40.1005 0 0.1 0.316583 63 0.00105
0.990028 0.009965 325.6131 0 0.01 2.437186 485 0.008083
0.999001 9.99E-04 2280.116 0 0.001 14.35678 2857 0.047617
0.9999 1.00E-04 15835.48 0 1.00E-04 91.31658 18172 0.302867
0.99991 9.00E-05 17342.06 0 9.00E-05 92.51256 18411 0.30685
0.99992 8.00E-05 19208.12 0 8.00E-05 102.7085 20439 0.34065
0.99993 7.00E-05 21562.95 0 7.00E-05 117.6583 23429 0.390483
0.99994 6.00E-05 24676.82 0 6.00E-05 132.7186 26411 0.440183
0.99995 5.00E-05 28974.85 0 5.00E-05 155.8442 31013 0.516883
0.99996 4.00E-05 35289.97 0 4.00E-05 190.0251 37815 0.63025
0.99997 3.00E-05 45565.71 0 3.00E-05 244.0704 48570 0.8095
0.99998 2.00E-05 65556.1 0 2.00E-05 353.6633 70379 1.172983
0.99999 1.00E-05 123256.9 0 1.00E-05 662.1508 131768 2.196133
0.999991 9.00E-06 135849.9 0 9.00E-06 729.8794 145246 2.420767
0.999992 8.00E-06 151453.6 0 8.00E-06 811.5477 161498 2.691633
0.999993 7.00E-06 171409.3 0 7.00E-06 924.8392 184043 3.067383
0.999994 6.00E-06 197814.9 0 6.00E-06 1064.402 211816 3.530267
0.999995 5.00E-06 234464.1 0 5.00E-06 1265.804 251895 4.19825
0.999996 4.00E-06 288965.3 0 4.00E-06 1556.377 309720 5.162
0.999997 3.00E-06 378816.9 0 3.00E-06 2045.628 407080 6.784667
0.999998 2.00E-06 556550.7 0 2.00E-06 2993.387 595699 9.928317
0.999999 1.00E-06 1080983 0 1.00E-06 5829.889 1160148 19.3358
0.999999 9.00E-07 1208134 0 9.00E-07 6812.869 1355770 22.59617
0.999999 8.00E-07 1353904 0 8.00E-07 7593.246 1511063 25.18438
0.999999 7.00E-07 1539773 0 7.00E-07 8703.136 1731927 28.86545
0.999999 6.00E-07 1786507 0 6.00E-07 10142.74 2018416 33.64027
1 5.00E-07 2133555 0 5.00E-07 10983.01 2185627 36.42712
1 4.00E-07 2646546 0 4.00E-07 14397.07 2865023 47.75038
1 3.00E-07 3500986 0 3.00E-07 18074.47 3596836 59.94727
1 2.00E-07 5199169 0 2.00E-07 28020.73 5576127 92.93545
1 1.00E-07 1.03E+07 0 1.00E-07 55535.69 11051602 184.1934
595.8261
d0.1*2*10*2
f2[0] f2[1] 迭代次数n 平均准确率p-ave δ 耗时ms/次 耗时ms/199次 耗时 min/199
0.525464 0.470535 1.834171 0 0.5 0.472362 94 0.001567
0.621887 0.377717 5.467337 0 0.4 0.311558 62 0.001033
0.713341 0.285849 10.37186 0 0.3 0.236181 47 0.000783
0.806947 0.193202 18.98492 0 0.2 0.236181 47 0.000783
0.901937 0.098125 41.81407 0 0.1 0.155779 47 0.000783
0.99002 0.009976 405.9146 0 0.01 1.648241 328 0.005467
0.999 1.00E-03 3909.161 0 0.001 12.82412 2552 0.042533
0.9999 1.00E-04 38025.66 0 1.00E-04 102.9799 20493 0.34155
0.99991 9.00E-05 42192.34 0 9.00E-05 113.4623 22579 0.376317
0.99992 8.00E-05 47395.7 0 8.00E-05 126.7688 25227 0.42045
0.99993 7.00E-05 54052.65 0 7.00E-05 144.0653 28669 0.477817
0.99994 6.00E-05 62950.38 0 6.00E-05 167.3769 33308 0.555133
0.99995 5.00E-05 75340.26 0 5.00E-05 200.1508 39830 0.663833
0.99996 4.00E-05 93917.49 0 4.00E-05 250.7136 49892 0.831533
0.99997 3.00E-05 124764.8 0 3.00E-05 333.8342 66433 1.107217
0.99998 2.00E-05 186176.6 0 2.00E-05 494.1055 98343 1.63905
0.99999 1.00E-05 368710.9 0 1.00E-05 981.5628 195331 3.255517
0.999991 9.00E-06 408984.1 0 9.00E-06 1096.543 218214 3.6369
0.999992 8.00E-06 459388.5 0 8.00E-06 1233.422 245451 4.09085
0.999993 7.00E-06 524038.9 0 7.00E-06 1400.101 278620 4.643667
0.999994 6.00E-06 609813.1 0 6.00E-06 1633.131 324993 5.41655
0.999995 5.00E-06 729919.3 0 5.00E-06 1953.291 388705 6.478417
0.999996 4.00E-06 909130.6 0 4.00E-06 2433.729 484312 8.071867
0.999997 3.00E-06 1207125 0 3.00E-06 3293.899 655489 10.92482
0.999998 2.00E-06 1799196 0 2.00E-06 4926.965 980469 16.34115
0.999999 1.00E-06 3559300 0 1.00E-06 10112.95 2012481 33.54135
0.999999 9.00E-07 3947485 0 9.00E-07 11262.18 2241184 37.35307
0.999999 8.00E-07 4433662 0 8.00E-07 11642.25 2316810 38.6135
0.999999 7.00E-07 5056312 0 7.00E-07 13749.6 2736171 45.60285
0.999999 6.00E-07 5882920 0 6.00E-07 15936.78 3171436 52.85727
1 5.00E-07 7040337 0 5.00E-07 18808.06 3742809 62.38015
1 4.00E-07 8766716 0 4.00E-07 23278.54 4632436 77.20727
1 3.00E-07 1.16E+07 0 3.00E-07 32613 6489996 108.1666
1 2.00E-07 1.73E+07 0 2.00E-07 47493.48 9451223 157.5204
1 1.00E-07 3.42E+07 0 1.00E-07 95195.08 18943821 315.7304
998.2984
d0.1-3-10-3-10-3
f2[0] f2[1] f2[2] 迭代次数n 平均准确率p-ave δ 耗时ms/次 耗时ms/199次 耗时 min/199
0.527193 0.470014 0.469798 1.959799 0 0.5 0.472362 94 0.001567
0.619754 0.378289 0.380506 5.341709 0 0.4 0.155779 31 0.000517
0.714597 0.286142 0.286613 10.1809 0 0.3 0.236181 47 0.000783
0.808487 0.192633 0.192337 18.52764 0 0.2 0.316583 63 0.00105
0.902196 0.097576 0.097796 38.36683 0 0.1 0.386935 93 0.00155
0.990043 0.009961 0.009955 268.5025 0 0.01 2.276382 453 0.00755
0.999001 9.99E-04 9.99E-04 1684.382 0 0.001 15.29146 3058 0.050967
0.9999 1.00E-04 1.00E-04 12056.82 0 1.00E-04 74.60804 14862 0.2477
0.99991 9.00E-05 9.00E-05 13253.44 0 9.00E-05 78.15075 15552 0.2592
0.99992 8.00E-05 8.00E-05 14739.48 0 8.00E-05 86.64322 17242 0.287367
0.99993 7.00E-05 7.00E-05 16633.47 0 7.00E-05 98.61809 19625 0.327083
0.99994 6.00E-05 6.00E-05 19138.09 0 6.00E-05 113.3417 22555 0.375917
0.99995 5.00E-05 5.00E-05 22613.09 0 5.00E-05 132.2714 26322 0.4387
0.99996 4.00E-05 4.00E-05 27783.5 0 4.00E-05 162.804 32398 0.539967
0.99997 3.00E-05 3.00E-05 36295.21 0 3.00E-05 212.5578 42299 0.704983
0.99998 2.00E-05 2.00E-05 53122.44 0 2.00E-05 311.206 61930 1.032167
0.99999 1.00E-05 1.00E-05 102725.2 0 1.00E-05 604.1658 120229 2.003817
0.999991 9.00E-06 9.00E-06 113718.9 0 9.00E-06 669.7688 133299 2.22165
0.999992 8.00E-06 8.00E-06 127307.4 0 8.00E-06 749.5025 149167 2.486117
0.999993 7.00E-06 7.00E-06 144868.4 0 7.00E-06 854.5678 170059 2.834317
0.999994 6.00E-06 6.00E-06 168071.1 0 6.00E-06 990.9598 197201 3.286683
0.999995 5.00E-06 5.00E-06 200668.7 0 5.00E-06 1182.101 235254 3.9209
0.999996 4.00E-06 4.00E-06 249317.9 0 4.00E-06 1471.251 292795 4.879917
0.999997 3.00E-06 3.00E-06 330180.3 0 3.00E-06 1945.874 387244 6.454067
0.999998 2.00E-06 2.00E-06 491437.5 0 2.00E-06 2906.714 578436 9.6406
0.999999 1.00E-06 1.00E-06 974155.5 0 1.00E-06 5805.045 1155204 19.2534
0.999999 9.00E-07 9.00E-07 1080634 0 9.00E-07 6453.688 1284299 21.40498
0.999999 8.00E-07 8.00E-07 1215192 0 8.00E-07 7235.307 1439826 23.9971
0.999999 7.00E-07 7.00E-07 1387027 0 7.00E-07 8266.302 1644994 27.41657
0.999999 6.00E-07 6.00E-07 1616168 0 6.00E-07 9634.271 1917251 31.95418
1 5.00E-07 5.00E-07 1937819 0 5.00E-07 11562.39 2300915 38.34858
1 4.00E-07 4.00E-07 2418742 0 4.00E-07 14425.52 2870695 47.84492
1 3.00E-07 3.00E-07 3222108 0 3.00E-07 20229.73 4025728 67.09547
1 2.00E-07 2.00E-07 4828570 0 2.00E-07 29352.16 5841084 97.3514
1 1.00E-07 1.00E-07 9653872 0 1.00E-07 58702.82 11681878 194.698
611.3697
d0.1-3-10-3
f2[0] f2[1] f2[2] 迭代次数n 平均准确率p-ave δ 耗时ms/次 耗时ms/199次 耗时 min/199
0.53421 0.464075 0.465333 2.150754 0 0.5 0.316583 78 0.0013
0.62713 0.374943 0.373477 5.668342 0 0.4 0.080402 32 0.000533
0.716243 0.282829 0.283456 10.52764 0 0.3 0.155779 31 0.000517
0.807594 0.191996 0.191729 19.09548 0 0.2 0.236181 47 0.000783
0.902494 0.097579 0.097602 41.92965 0 0.1 0.236181 47 0.000783
0.990029 0.009972 0.009973 400.9246 0 0.01 1.884422 375 0.00625
0.999 1.00E-03 1.00E-03 3785.151 0 0.001 12.9799 2599 0.043317
0.9999 1.00E-04 1.00E-04 35887.74 0 1.00E-04 107.5578 21404 0.356733
0.99991 9.00E-05 9.00E-05 39777.31 0 9.00E-05 -216.437 -43071 -0.71785
0.99992 8.00E-05 8.00E-05 44621.81 0 8.00E-05 129.2462 25735 0.428917
0.99993 7.00E-05 7.00E-05 50836.96 0 7.00E-05 149.8442 29819 0.496983
0.99994 6.00E-05 6.00E-05 59079.19 0 6.00E-05 174.3869 34703 0.578383
0.99995 5.00E-05 5.00E-05 70575.01 0 5.00E-05 207.7035 41333 0.688883
0.99996 4.00E-05 4.00E-05 87725.12 0 4.00E-05 262.7387 52316 0.871933
0.99997 3.00E-05 3.00E-05 116155.5 0 3.00E-05 346.3065 68915 1.148583
0.99998 2.00E-05 2.00E-05 172367 0 2.00E-05 508.2111 101134 1.685567
0.99999 1.00E-05 1.00E-05 338333.8 0 1.00E-05 1003.648 199726 3.328767
0.999991 9.00E-06 9.00E-06 375023.1 0 9.00E-06 1112.487 221385 3.68975
0.999992 8.00E-06 8.00E-06 420561.9 0 8.00E-06 1250.854 248920 4.148667
0.999993 7.00E-06 7.00E-06 478899.1 0 7.00E-06 1418.658 282313 4.705217
0.999994 6.00E-06 6.00E-06 556314.6 0 6.00E-06 1648.251 328017 5.46695
0.999995 5.00E-06 5.00E-06 664199.1 0 5.00E-06 1971.07 392275 6.537917
0.999996 4.00E-06 4.00E-06 825530.8 0 4.00E-06 2453.859 488318 8.138633
0.999997 3.00E-06 3.00E-06 1091743 0 3.00E-06 3247.789 646310 10.77183
0.999998 2.00E-06 2.00E-06 1619069 0 2.00E-06 4810.317 957268 15.95447
0.999999 1.00E-06 1.00E-06 3175825 0 1.00E-06 9445.754 1879705 31.32842
0.999999 9.00E-07 9.00E-07 3517441 0 9.00E-07 10347.84 2059237 34.32062
0.999999 8.00E-07 8.00E-07 3943425 0 8.00E-07 11567.7 2301973 38.36622
0.999999 7.00E-07 7.00E-07 4490016 0 7.00E-07 13167.86 2620404 43.6734
0.999999 6.00E-07 6.00E-07 5216202 0 6.00E-07 15299.34 3044568 50.7428
1 5.00E-07 5.00E-07 6225307 0 5.00E-07 18243.98 3630569 60.50948
1 4.00E-07 4.00E-07 7731842 0 4.00E-07 22756.38 4528536 75.4756
1 3.00E-07 3.00E-07 1.02E+07 0 3.00E-07 31021.67 6173330 102.8888
1 2.00E-07 2.00E-07 1.52E+07 0 2.00E-07 45639.92 9082360 151.3727
1 1.00E-07 1.00E-07 2.97E+07 0 1.00E-07 88688.07 17648925 294.1488
951.1606
d0.1-4-10-4-10-4
f2[0] f2[1] f2[2] f2[3] 迭代次数n 平均准确率p-ave δ 耗时ms/次 耗时ms/199次 耗时 min/199
0.525973 0.470305 0.472377 0.472429 3 0 0.5 0.547739 109 0.001817
0.622486 0.377597 0.379318 0.375209 7 0 0.4 0.160804 32 0.000533
0.713703 0.285557 0.286032 0.286273 11 0 0.3 0.311558 62 0.001033
0.807452 0.192672 0.192159 0.192303 20 0 0.2 0.316583 63 0.00105
0.902426 0.097368 0.097147 0.097417 39 0 0.1 0.547739 109 0.001817
0.990052 0.009949 0.009942 0.009945 237 0 0.01 2.201005 438 0.0073
0.999001 9.99E-04 9.99E-04 9.99E-04 1406 0 0.001 10.94472 2210 0.036833
0.9999 9.99E-05 9.99E-05 9.99E-05 10609 0 1.00E-04 72.35176 14415 0.24025
0.99991 9.00E-05 9.00E-05 9.00E-05 11717 0 9.00E-05 77.26131 15385 0.256417
0.99992 8.00E-05 8.00E-05 8.00E-05 13077 0 8.00E-05 84.24121 16768 0.279467
0.99993 7.00E-05 7.00E-05 7.00E-05 14844 0 7.00E-05 102.5628 20433 0.34055
0.99994 6.00E-05 6.00E-05 6.00E-05 17197 0 6.00E-05 116.2764 23143 0.385717
0.99995 5.00E-05 5.00E-05 5.00E-05 20480 0 5.00E-05 136.5075 27185 0.453083
0.99996 4.00E-05 4.00E-05 4.00E-05 25306 0 4.00E-05 169.9799 33831 0.56385
0.99997 3.00E-05 3.00E-05 3.00E-05 33318 0 3.00E-05 226.7538 45131 0.752183
0.99998 2.00E-05 2.00E-05 2.00E-05 49468 0 2.00E-05 334.9296 66656 1.110933
0.99999 1.00E-05 1.00E-05 9.99E-06 97258 0 1.00E-05 661.0955 131561 2.192683
0.999991 9.00E-06 9.00E-06 9.00E-06 108216 0 9.00E-06 731.598 145588 2.426467
0.999992 8.00E-06 8.00E-06 8.00E-06 121518 0 8.00E-06 799.6834 159143 2.652383
0.999993 7.00E-06 7.00E-06 7.00E-06 139247 0 7.00E-06 917.608 182612 3.043533
0.999994 6.00E-06 6.00E-06 6.00E-06 161784 0 6.00E-06 1062.995 211538 3.525633
0.999995 5.00E-06 5.00E-06 5.00E-06 193666 0 5.00E-06 1273.06 253343 4.222383
0.999996 4.00E-06 4.00E-06 4.00E-06 241691 0 4.00E-06 1589.93 316400 5.273333
0.999997 3.00E-06 3.00E-06 3.00E-06 322513 0 3.00E-06 2151.678 428189 7.136483
0.999998 2.00E-06 2.00E-06 2.00E-06 482398 0 2.00E-06 3193.673 635560 10.59267
0.999999 1.00E-06 1.00E-06 1.00E-06 964074 0 1.00E-06 6390.251 1271663 21.19438
0.999999 9.00E-07 9.00E-07 9.00E-07 1072751 0 9.00E-07 7089.07 1410734 23.51223
0.999999 8.00E-07 8.00E-07 8.00E-07 1206559 0 8.00E-07 7998.598 1591723 26.52872
0.999999 7.00E-07 7.00E-07 7.00E-07 1381749 0 7.00E-07 8947.251 1780505 29.67508
0.999999 6.00E-07 6.00E-07 6.00E-07 1606431 0 6.00E-07 10406.81 2070956 34.51593
1 5.00E-07 5.00E-07 5.00E-07 1931582 0 5.00E-07 12522.17 2491915 41.53192
1 4.00E-07 4.00E-07 4.00E-07 2414921 0 4.00E-07 15727.61 3129795 52.16325
1 3.00E-07 3.00E-07 3.00E-07 3226161 0 3.00E-07 20389.41 4057494 67.6249
1 2.00E-07 2.00E-07 2.00E-07 4839273 0 2.00E-07 31133.98 6195679 103.2613
1 1.00E-07 1.00E-07 1.00E-07 9700878 0 1.00E-07 63021.79 12541354 209.0226
654.5287
d0.1-4-10-4
f2[0] f2[1] f2[2] f2[3] 迭代次数n 平均准确率p-ave δ 耗时ms/次 耗时ms/199次 耗时 min/199
0.536878 0.460988 0.46149 0.460007 4 0 0.5 0.557789 111 0.00185
0.629137 0.371521 0.3726 0.372414 7 0 0.4 0.155779 31 0.000517
0.7184 0.281781 0.280531 0.280074 12 0 0.3 0.100503 35 0.000583
0.809433 0.191093 0.190154 0.190257 21 0 0.2 0.236181 47 0.000783
0.902386 0.097307 0.097211 0.097206 43 0 0.1 0.236181 47 0.000783
0.990031 0.009969 0.009966 0.009967 397 0 0.01 2 415 0.006917
0.999 1.00E-03 1.00E-03 1.00E-03 3660 0 0.001 15.26633 3038 0.050633
0.9999 1.00E-04 1.00E-04 1.00E-04 33662 0 1.00E-04 115.1357 22912 0.381867
0.99991 9.00E-05 9.00E-05 9.00E-05 37338 0 9.00E-05 126.6784 25225 0.420417
0.99992 8.00E-05 8.00E-05 8.00E-05 41802 0 8.00E-05 144.9698 28865 0.481083
0.99993 7.00E-05 7.00E-05 7.00E-05 47443 0 7.00E-05 166.5276 33155 0.552583
0.99994 6.00E-05 6.00E-05 6.00E-05 55259 0 6.00E-05 190.598 37929 0.63215
0.99995 5.00E-05 5.00E-05 5.00E-05 65584 0 5.00E-05 227.9196 45361 0.756017
0.99996 4.00E-05 4.00E-05 4.00E-05 81242 0 4.00E-05 283.6382 56453 0.940883
0.99997 3.00E-05 3.00E-05 3.00E-05 107253 0 3.00E-05 376.0503 74842 1.247367
0.99998 2.00E-05 2.00E-05 2.00E-05 158480 0 2.00E-05 554.1608 110278 1.837967
0.99999 1.00E-05 1.00E-05 1.00E-05 308464 0 1.00E-05 1076.221 214177 3.569617
0.999991 9.00E-06 9.00E-06 9.00E-06 341247 0 9.00E-06 1192.693 237353 3.955883
0.999992 8.00E-06 8.00E-06 8.00E-06 382394 0 8.00E-06 1328.07 264294 4.4049
0.999993 7.00E-06 7.00E-06 7.00E-06 435268 0 7.00E-06 1515.171 301527 5.02545
0.999994 6.00E-06 6.00E-06 6.00E-06 504206 0 6.00E-06 1757.176 349694 5.828233
0.999995 5.00E-06 5.00E-06 5.00E-06 600920 0 5.00E-06 2055.226 409007 6.816783
0.999996 4.00E-06 4.00E-06 4.00E-06 744853 0 4.00E-06 2519.915 501466 8.357767
0.999997 3.00E-06 3.00E-06 3.00E-06 980381 0 3.00E-06 3350.477 666746 11.11243
0.999998 2.00E-06 2.00E-06 2.00E-06 1449669 0 2.00E-06 4943.613 983782 16.39637
0.999999 1.00E-06 1.00E-06 1.00E-06 2818734 0 1.00E-06 9529.804 1896447 31.60745
0.999999 9.00E-07 9.00E-07 9.00E-07 3123123 0 9.00E-07 10445.02 2078560 34.64267
0.999999 8.00E-07 8.00E-07 8.00E-07 3493458 0 8.00E-07 11695.43 2327390 38.78983
0.999999 7.00E-07 7.00E-07 7.00E-07 3972648 0 7.00E-07 13545.04 2695464 44.9244
0.999999 6.00E-07 6.00E-07 6.00E-07 4605197 0 6.00E-07 15193.38 3023483 50.39138
1 5.00E-07 5.00E-07 5.00E-07 5492530 0 5.00E-07 18164.42 3614719 60.24532
1 4.00E-07 4.00E-07 4.00E-07 6800941 0 4.00E-07 22708.23 4518943 75.31572
1 3.00E-07 3.00E-07 3.00E-07 8973350 0 3.00E-07 30762.9 6121852 102.0309
1 2.00E-07 2.00E-07 2.00E-07 1.32E+07 0 2.00E-07 44557.35 8866912 147.7819
1 1.00E-07 1.00E-07 1.00E-07 2.58E+07 0 1.00E-07 88140.6 17539982 292.333
950.8424
本文原始数据有感兴趣的朋友可以到我的资源里下载