到底应该加几个卷积核?

用81-30-2的网络来二分类mnist的0和1，然后再向这个网络上加卷积核，如果想要性能和效率都实现最优应该加几个最合适？

 

首先做对照网络

d2（mnist 0,1）81-30-2-（2*k） ,k∈{0,1}

用81-30-2的网络来二分类mnist的0和1，并让0向（1，0）收敛，让1向（0，1）收敛。

用这个没有加卷积核的网络的性能做参照来比较增加卷积核对网络性能的影响。

再制作有卷积核的网络

d2（mnist 0,1）81-con（3*3*n）-（49*n）-30-2-（2*k） ,k∈(0,1)

用带有n个3*3卷积核的网络来分类mnist的0和1，网络结构是（49*n）-30-2. 并让0向（1，0）收敛，让1向（0，1）收敛。让卷积核的数量n分别等于1到9.

也就是共制作了10个网络，让网络的收敛标准统一为

if (Math.abs(网路输出值[0]-目标函数[0])< δ  &&  Math.abs(网络输出值[1]-目标函数[1])< δ   )，

其中δ分别等于0.5到1e-6.的34个值。对应每个δ收敛199次，分别记录与之对应的迭代次数，收敛时间，并计算199次的平均准确率和199次的最大准确率来比较这10个网络的性能差异。共收敛了34*199*10次

首先比较这10个网络的最大分辨准确率

 

	81-30-2	N=1	2	3	4	5	6	7	8	9
δ	最大值p-max	最大值p-max	最大值p-max	最大值p-max	最大值p-max	最大值p-max	最大值p-max	最大值p-max	最大值p-max	最大值p-max
0.5	0.9881797	0.9054374	0.9687943	0.9579196	0.9546099	0.9347518	0.9867612	0.9706856	0.9820331	0.9910165
0.4	0.9981087	0.9971631	0.9943262	0.9981087	0.9976359	0.9962175	0.9962175	0.9966903	0.9981087	0.9966903
0.3	0.9981087	0.9981087	0.9957447	0.9971631	0.9981087	0.9976359	0.9981087	0.9981087	0.9981087	0.9981087
0.2	0.9990544	0.9981087	0.9966903	0.9976359	0.9981087	0.9985816	0.9976359	0.9985816	0.9981087	0.9976359
0.1	0.9995272	0.9981087	0.9957447	0.9971631	0.9976359	0.9976359	0.9976359	0.9981087	0.9981087	0.9985816
0.01	0.9990544	0.9971631	0.9971631	0.9981087	0.9981087	0.9985816	0.9990544	0.9990544	0.9995272	0.9985816
0.001	0.9990544	0.9985816	0.9990544	0.9990544	0.9985816	0.9985816	0.9990544	0.9990544	0.9990544	0.9985816
9.00E-04	0.9990544	0.9981087	0.9981087	0.9990544	0.9985816	0.9990544	0.9995272	0.9990544	0.9990544	0.9990544
8.00E-04	0.9990544	0.9985816	0.9985816	0.9990544	0.9990544	0.9990544	0.9990544	0.9990544	0.9995272	0.9990544
7.00E-04	0.9990544	0.9985816	0.9990544	0.9990544	0.9995272	0.9990544	0.9995272	0.9995272	0.9995272	0.9990544
6.00E-04	0.9990544	0.9981087	0.9985816	0.9990544	0.9995272	0.9995272	0.9995272	0.9995272	0.9995272	0.9995272
5.00E-04	0.9990544	0.9981087	0.9990544	0.9995272	0.9995272	0.9995272	0.9995272	0.9995272	0.9995272	0.9995272
4.00E-04	1	0.9985816	0.9995272	1	0.9995272	0.9995272	0.9995272	0.9995272	0.9995272	0.9995272
3.00E-04	1	0.9995272	0.9995272	1	1	0.9995272	1	1	0.9995272	1
2.00E-04	1	0.9985816	0.9995272	1	1	1	1	1	1	1
1.00E-04	0.9990544	0.9985816	1	1	1	1	1	1	1	1
9.00E-05	0.9990544	0.9995272	1	1	1	1	1	1	1	1
8.00E-05	0.9995272	0.9985816	1	1	1	1	1	1	1	1
7.00E-05	0.9995272	0.9995272	1	1	1	1	1	1	1	1
6.00E-05	0.9995272	0.9990544	1	1	1	1	1	1	1	1
5.00E-05	0.9995272	0.9990544	1	1	1	1	1	1	1	1
4.00E-05	0.9995272	0.9990544	1	1	1	1	1	1	1	1
3.00E-05	0.9995272	0.9990544	1	1	1	1	1	1	1	1
2.00E-05	0.9995272	0.9990544	1	1	1	1	1	1	1	1
1.00E-05	0.9995272	1	1	1	1	1	1	1	1	1
9.00E-06	0.9995272	0.9995272	1	1	1	1	1	1	1	1
8.00E-06	0.9995272	0.9990544	1	1	1	1	1	1	1	1
7.00E-06	0.9995272	0.9995272	1	1	1	1	1	1	1	1
6.00E-06	0.9995272	0.9995272	1	1	1	1	1	1	1	1
5.00E-06	1	0.9995272	1	1	1	1	1	1	1	1
4.00E-06	1	1	1	1	1	1	1	1	1	1
3.00E-06	1	1	1	1	1	1	1	1	1	1
2.00E-06	1	0.9995272	1	1	1	1	1	1	1	1
1.00E-06	1	1	1	1	1	1	1	1	1	1

效果是很显著的当卷积核的数量>=2的时候网络的最大准确率当δ<=1e-4的时候已经为1，要显著的强于未加卷积核的81-30-2的网络。或许是由于测试集的难度太小了，只观察到了2个卷积核的最大准确率要好于1个卷积核的网络和当δ=0.5是最大准确率随着卷积核的数量的增加而增加，除此并未发现卷积核数量的增加会改善网络最大分辨准确率的现象。

再比较平均分辨准确率

	81-30-2	1	2	3	4	5	6	7	8	9
δ	平均准确率p-ave	平均准确率p-ave	平均准确率p-ave	平均准确率p-ave	平均准确率p-ave	平均准确率p-ave	平均准确率p-ave	平均准确率p-ave	平均准确率p-ave	平均准确率p-ave
0.5	0.5751737	0.503242	0.5279756	0.5106597	0.5228435	0.5252337	0.5275835	0.5127529	0.5360657	0.5396533
0.4	0.9954406	0.7974934	0.8965323	0.8834884	0.8674745	0.830127	0.8256673	0.7895553	0.7497986	0.7164261
0.3	0.9952647	0.9536002	0.9722537	0.9794124	0.9801205	0.9772218	0.9764377	0.9727384	0.9708923	0.9678321
0.2	0.996429	0.9703102	0.9787139	0.9882129	0.989111	0.988346	0.9892797	0.9897692	0.9887166	0.9880086
0.1	0.9974815	0.9784288	0.9860579	0.9907267	0.9918505	0.9929838	0.9927676	0.9933877	0.9934923	0.9935612
0.01	0.9986671	0.9780344	0.9859035	0.9908051	0.9919075	0.9912541	0.9875833	0.9876451	0.9869275	0.986153
0.001	0.9987479	0.9844185	0.9938677	0.9944712	0.9857206	0.9875144	0.9798805	0.9729427	0.9615857	0.9507158
9.00E-04	0.9986932	0.9881892	0.9954358	0.9955712	0.9923827	0.9928199	0.9906174	0.9851456	0.9808736	0.9716859
8.00E-04	0.9985008	0.9868729	0.9960464	0.9959229	0.9945591	0.9949749	0.9936871	0.9924754	0.9904677	0.9890968
7.00E-04	0.9981776	0.9881963	0.9964337	0.9960678	0.9954382	0.995652	0.9952196	0.9944926	0.993181	0.9939485
6.00E-04	0.9984913	0.9829051	0.996581	0.9965382	0.9965668	0.9965501	0.9961557	0.9956924	0.9949012	0.9934329
5.00E-04	0.9987122	0.9787567	0.9969445	0.9970895	0.9969968	0.9969659	0.9968946	0.9955831	0.9960298	0.9952885
4.00E-04	0.998477	0.9836107	0.9975053	0.9975456	0.9975361	0.9974886	0.996859	0.9967568	0.9959609	0.9944545
3.00E-04	0.9993371	0.9803295	0.9977167	0.9977714	0.9972795	0.9968875	0.9966333	0.9953431	0.9648479	0.9940578
2.00E-04	0.9993324	0.9759863	0.9977476	0.9974791	0.9966879	0.9959823	0.9947088	0.993775	0.9880775	0.9921261
1.00E-04	0.9978474	0.9850553	0.9984414	0.9978403	0.9976027	0.9964884	0.9957043	0.9951008	0.9960345	0.9955356
9.00E-05	0.9980208	0.9857229	0.998477	0.9984865	0.9981824	0.9969374	0.9971251	0.9963102	0.9965929	0.9952671
8.00E-05	0.9983939	0.9856612	0.9986647	0.9982085	0.9984438	0.9974839	0.9974934	0.9966309	0.9974839	0.9960916
7.00E-05	0.9986861	0.9855542	0.9987479	0.9984105	0.9985079	0.9979709	0.9977975	0.9975076	0.9975932	0.9964456
6.00E-05	0.9989475	0.9819452	0.9987503	0.9987431	0.9987051	0.998439	0.9979923	0.9976526	0.998047	0.9972415
5.00E-05	0.999406	0.9837557	0.9988667	0.9987241	0.9986006	0.9983891	0.998161	0.9980517	0.9978973	0.9973271
4.00E-05	0.9995248	0.9804436	0.9987455	0.9989617	0.9986338	0.9955546	0.9982774	0.9980256	0.9980826	0.9975789
3.00E-05	0.9994773	0.9804673	0.9989071	0.9988144	0.9986932	0.9980969	0.9982632	0.9982299	0.9983345	0.9979804
2.00E-05	0.9993775	0.9848581	0.9988643	0.9989736	0.9987146	0.9982561	0.99842	0.9982513	0.9983558	0.9982988
1.00E-05	0.9993181	0.9863383	0.9989498	0.9989142	0.9986837	0.9980612	0.9983487	0.99866	0.9985317	0.9982062
9.00E-06	0.9992991	0.9874716	0.9989641	0.9987954	0.9986481	0.9984509	0.998363	0.9984889	0.9983036	0.9974554
8.00E-06	0.9992373	0.9874503	0.9990401	0.9963648	0.9988001	0.9983962	0.9962721	0.9986315	0.996745	0.9980232
7.00E-06	0.9992326	0.9876688	0.9989498	0.9985032	0.9987479	0.9974744	0.9980446	0.9985602	0.9974435	0.9979709
6.00E-06	0.9992848	0.9870345	0.9988833	0.9985222	0.9987289	0.9979353	0.9980541	0.9986837	0.998382	0.9981872
5.00E-06	0.9994155	0.9861554	0.998957	0.9985032	0.9988144	0.9984461	0.9983749	0.9986124	0.9980208	0.9982537
4.00E-06	0.9995153	0.9856065	0.9989641	0.998553	0.9987289	0.9983273	0.9987004	0.9957756	0.9982727	0.9982584
3.00E-06	0.9996175	0.987733	0.9989261	0.9987669	0.9976146	0.9986766	0.99866	0.9975361	0.996859	0.9983202
2.00E-06	0.9997933	0.9902491	0.998938	0.9988049	0.9980921	0.9987027	0.99866	0.9980493	0.9970348	0.9965858
1.00E-06	0.9995985	0.9919836	0.9989926	0.9987312	0.9984485	0.9969303	0.9967545	0.9982988	0.997586	0.9980113

这个结果是很意外的，对本次实验的1-9个卷积核，无论加多少个卷积核网络的平均性能都不如未加卷积核的对照网络。也就是如果更在乎平均性能不如不加卷积核。增加卷积核对81*30*2的网络的平均性能没有任何正面影响，而且随着卷积核数量的增加平均性能波动越大，越不稳定。

 

第三比较迭代次数

	81-30-2	1	2	3	4	5	6	7	8	9
δ	迭代次数n	迭代次数n	迭代次数n	迭代次数n	迭代次数n	迭代次数n	迭代次数n	迭代次数n	迭代次数n	迭代次数n
0.5	8.9849246	17.914573	18.356784	16.407035	17.482412	15.547739	17.467337	12.60804	14.236181	14.623116
0.4	151.46734	1272.9146	155.21106	135.27638	124.34171	117.74372	113.80905	110.52261	106.59799	106.54271
0.3	193.96482	1427.2714	99	84.81407	86.361809	85.241206	81.59799	84.592965	79.839196	80.693467
0.2	234.47739	1559.201	97.246231	92.437186	85.959799	87.366834	88.311558	85.276382	80.633166	85.889447
0.1	308.96482	1718.005	116.82915	116.30151	113.48241	115.65829	117.47739	124.9196	117.65829	117.50754
0.01	648.04523	2196.4422	298.80402	326.79397	338.24623	338.8392	339.51759	350.22111	347.95477	351.40201
0.001	1987.2211	3121.9749	1387.6583	1500.9146	1359.9397	1463.8492	1509.4573	1473.794	1451.7889	1397.1156
9.00E-04	2099.9447	3145.995	1521.3116	1674.1809	1741.0402	1752.8995	1745.9196	1758.8995	1784.8291	1689.8291
8.00E-04	2158.2161	3234.5729	1683.397	1929.9246	2128.1256	2020.1005	1998.7387	2082.3819	2139.2412	2169.9799
7.00E-04	2339.5528	3323.3719	1933.0452	2141.201	2405.0151	2250.4523	2339.8693	2384.5025	2390.9799	2421.4372
6.00E-04	2616.3719	3390.7889	2254.794	2411.7437	2881.5528	2584.4573	2677.9648	2801.3116	2781.1608	2799.4774
5.00E-04	2874.1709	3510.9196	2749.6181	2905.4774	3194.9397	3009.8894	3087.598	3136.9698	3252.9045	3256.0452
4.00E-04	3103.6784	3657.8693	3439.6583	3576.0704	4063.3116	3557.0352	3701.4874	3743.7688	3885.201	3781.995
3.00E-04	4253.8643	3734.407	4653.0854	4458.1709	5366.3015	4798.6784	5027.9849	5106.8995	3931.8844	5129.3216
2.00E-04	5159.8342	4094.7638	7342.8543	6998.9347	7857.2864	6804.7437	7065.5126	7001.5779	9039.8241	7125.9598
1.00E-04	5313.5477	4911.5729	15506.523	14384.759	15661.156	12383.598	13033.332	12325.96	18030.437	12689.749
9.00E-05	5457.6281	4893.2412	13203.487	14424.467	15587.477	14052.729	13926.789	13759.533	19101.362	13822.01
8.00E-05	5814.4121	5162.2563	13376.839	14449.583	15865.03	15632.03	15787.98	14914.91	19248.337	15260.698
7.00E-05	6418.995	5260.407	14084.156	15548.824	17521.437	17895.95	17362.663	17504.095	21427.548	17085.739
6.00E-05	7637.1055	5498.7236	15846.543	17738.523	18500.673	19363.005	20086.668	19431.678	23328.92	18479.136
5.00E-05	9614.9548	5636.9146	16720.432	19301.136	19821.92	21446.03	23328.317	21723.905	25601.452	22090.452
4.00E-05	12084.593	6084.8442	19631.955	20924.327	22956	31802.839	26592.92	25417.492	29984.538	25991.136
3.00E-05	14731.327	6630.5879	22636.035	23709.357	27418.879	37657.161	31323.106	31166.216	33767.291	32577.467
2.00E-05	15502.673	7487.6985	28829.749	31211.613	34490.317	39645.352	41799.608	39314.266	43519.91	39049.779
1.00E-05	23538.915	9588.9899	39033.603	43837.337	48120.171	53363.085	60223.447	73658.322	63562.337	62636.548
9.00E-06	24890.412	10059.462	36848.085	42105.221	46714.543	51339.673	60674.99	62954.739	65050.985	86731.849
8.00E-06	26143.367	10449.889	37868.141	51632.447	47212.638	51188.322	78991.548	64876.879	77089.065	90758.111
7.00E-06	26976.925	11049.116	37583.121	49078.015	49336.06	77134.93	62342.04	69872.558	66253.085	79820.975
6.00E-06	30039.93	11802.126	39620.457	47897.337	52680.538	68720.065	63644.176	77103.864	68074	75305.739
5.00E-06	33905.859	12674.02	43121.608	49984.935	52986.186	68404.271	65834.437	75845.513	75117.256	85185.03
4.00E-06	38340.221	15614.482	44738.422	53647.573	58730.332	70417.965	75075.065	88052.658	76259.136	99206.302
3.00E-06	44902.06	17613.286	49216.608	57838.588	80131.437	79051.045	82619.769	91937.769	97506.925	89498.804
2.00E-06	56018.352	24346.503	56637.367	67256.789	83585.859	87650.271	91820.995	97789.352	117858.88	125675.34
1.00E-06	72385.156	41275.352	70958.894	85642.864	101804.81	117332.54	131001.35	129437.56	143227.49	145569.59

迭代次数随着卷积核的数量的增加而增加，当δ=1e-6的时候81-30-2的迭代次数大于2个核小于3个核。

迭代次数与卷积核的数量之间有怎样的关系？

当δ=1e-6的时候

i	1	2	3	4	5	6	7	8	9
	41275.3518	70958.8945	85642.8643	101804.809	117332.543	131001.352	129437.558	143227.487	145569.588
n9/ni	3.52679218	2.05146358	1.69972816	1.42988911	1.24065826	1.11120676	1.12463176	1.01635231
(9/i)^0.5	3	2.12132034	1.73205081	1.5	1.34164079	1.22474487	1.13389342	1.06066017

比较n9/ni he (9/i)^0.5的曲线

如果不用1e-6的数据

		1	2	3	4	5	6	7	8
	(9/i)^0.5	3	2.12132034	1.73205081	1.5	1.34164079	1.22474487	1.13389342	1.06066017
1.00E-06	n9/ni	3.52679218	2.05146358	1.69972816	1.42988911	1.24065826	1.11120676	1.12463176	1.01635231
2.00E-06	n9/ni	5.16194622	2.21894738	1.8685896	1.50354782	1.43382713	1.36869936	1.28516382	1.06632043
3.00E-06	n9/ni	5.08132337	1.81846754	1.54738916	1.11690002	1.13216471	1.08326137	0.97347157	0.91787126
4.00E-06	n9/ni	6.35347999	2.21747431	1.84922255	1.6891834	1.40882091	1.32142811	1.12667015	1.30091038

可以发现δ越大数据偏离（9/i）^0.5的数据越大，因此可以得出δ越小应该偏离的越小

因此卷积核的数量和迭代次数的关系可以近似为

同一个网络分别用两个卷积核i和j，迭代次数ni与nj的比例与i和j的平方根近似成正比。

 

再比较收敛时间，

	81-30-2	1	2	3	4	5	6	7	8	9
δ	耗时 min/199	耗时 min/199	耗时 min/199	耗时 min/199	耗时 min/199	耗时 min/199	耗时 min/199	耗时 min/199	耗时 min/199	耗时 min/199
0.5	3.8609	2.9683333	3.22915	3.0758167	3.0727	2.6932	2.8877333	3.2802	3.8943667	3.1882667
0.4	3.9897833	5.1572667	3.7429167	3.00815	3.38625	2.8540667	3.0715167	3.5064667	4.2278	3.5390833
0.3	4.0389667	5.28525	3.5890333	2.93415	3.44655	2.7955833	3.0089167	3.4277	4.0827333	3.4944333
0.2	4.1132833	5.4824167	3.6008333	2.9411833	3.2820167	2.8023833	3.0064833	3.4240167	4.2887333	3.48205
0.1	4.2630833	5.7441667	3.6189333	2.9800833	3.2931667	2.8563	3.0642167	3.4796	4.4574	3.5777667
0.01	4.8257167	6.294	3.9175	3.30495	3.6747833	3.2552833	3.5282333	3.9865667	4.9973	4.1705667
0.001	7.1519167	7.4962667	5.52965	5.0543667	5.0211833	5.2103833	5.8818833	4.0525167	7.8104667	6.89225
9.00E-04	7.5316	7.6164333	5.6360167	5.3209833	5.6673333	5.7203833	6.35665	7.1377	8.3780833	7.6791833
8.00E-04	7.5574333	5.6474667	5.8822833	5.69605	6.27995	6.20195	3.4165333	7.8454667	9.2552167	8.5549667
7.00E-04	7.61015	7.8024833	6.3749333	6.01875	6.6235833	6.5990333	7.4027667	9.10465	9.9400333	9.1247667
6.00E-04	6.68905	7.9165667	4.7836	6.7381833	7.5925167	7.1919667	8.1009	10.543033	11.05375	6.6901833
5.00E-04	8.42065	7.97625	7.5541833	7.9375333	8.0095833	7.9301833	8.9317333	11.324817	9.0229667	11.2863
4.00E-04	8.9801167	8.3453	8.63575	8.6719167	9.36495	8.8807167	10.15705	12.484183	13.80275	12.459117
3.00E-04	11.11055	8.44955	10.363067	9.9560333	8.6762	11.110583	12.570967	13.629883	12.0492	15.778883
2.00E-04	12.5902	8.8268	14.571683	13.782333	15.49335	14.574483	13.937267	19.829783	23.303883	19.0359
1.00E-04	11.362533	7.459	24.315283	20.883117	28.114617	22.70635	27.256617	32.459683	43.592283	34.79835
9.00E-05	12.6392	10.035217	23.01875	24.37365	25.660867	27.108233	29.82835	38.233017	44.162883	34.761867
8.00E-05	13.284083	10.18275	23.097533	22.7186	28.533333	29.837433	33.84525	36.271417	46.440783	38.993517
7.00E-05	14.2286	10.175667	22.748633	26.372533	28.32705	35.625617	34.566933	43.42285	50.8357	45.915383
6.00E-05	18.913833	10.482533	27.4144	27.024	32.6756	33.28815	39.285333	47.61595	53.667883	46.710967
5.00E-05	22.811917	10.659833	26.028067	32.1013	31.331467	39.697517	48.636317	53.922	60.458567	56.073933
4.00E-05	28.900717	9.4542833	33.178667	31.454467	37.158517	57.390517	53.062383	61.338983	69.37795	64.793833
3.00E-05	34.315117	12.241883	33.858083	37.593033	47.695583	65.282	60.853517	74.435183	76.400183	81.527783
2.00E-05	35.70935	13.391867	45.3707	47.2158	55.3937	70.173383	80.77465	93.8695	101.82857	96.999667
1.00E-05	49.970433	16.2405	59.216133	67.597417	73.8599	96.508467	117.40512	154.8208	143.13805	163.38248
9.00E-06	54.454267	14.147383	56.49225	66.823733	75.700417	89.18165	116.0364	132.55487	147.19133	219.18537
8.00E-06	58.624667	17.571167	54.479133	74.137283	75.429417	89.8726	152.68102	149.9558	175.4566	226.97932
7.00E-06	63.180267	18.347517	50.881683	70.2156	77.116383	135.361	119.15367	160.45435	151.73543	199.74935
6.00E-06	68.21255	15.984233	55.734583	68.86575	82.778017	118.64055	121.53742	173.64287	157.87352	191.23247
5.00E-06	78.435767	20.081167	58.66405	70.860867	86.4666	121.92053	126.91468	171.95883	182.70695	209.22285
4.00E-06	85.330317	23.638433	57.240867	74.1	90.572867	140.49517	143.54072	184.00032	194.00147	252.12315
3.00E-06	99.625517	25.88465	65.804133	81.871683	126.52058	141.22632	157.5286	190.83585	225.15483	240.72125
2.00E-06	123.47987	34.638367	76.559983	93.11105	132.01262	159.31823	175.38062	224.37283	288.27608	333.47802
1.00E-06	158.16015	56.345083	94.149567	117.80548	155.58073	205.7136	254.05877	296.57368	356.3124	413.68133

随着卷积核数量的增加收敛时间也是增加的。

 

总结这4个表格

1.增加卷积核可以改善网络的最大性能
2.增加卷积核会降低网络的平均性能，而且卷积核数量越多平均性能越不稳定
3.同一个网络在收敛标准足够小的情况下迭代次数的比与卷积核的数量的平方根的比成正比 
 
4.卷积核的数量越多收敛速度越慢

将这10个网络的性能排序

最大性能	9≈	8≈	7≈	6≈	5≈	4≈	3≈	2>	1≈	81-30-2
平均性能	81-30-2>	1>	2>	3>	4>	5>	6>	7>	8>	9
迭代次数	9>	8>	7>	6>	5>	4>	3>	81-30-2>	2>	1

因此81-30-2的网络的最大性能小于卷积核数量大于2的网络，平均性能大于任何加卷积核的网络，迭代次数在2个核与3个核之间，所以针对实验的网络和1-9个核，如果更在乎平均性能就不加卷积核，如果在乎最大性能，最合理的卷积核的数量就是2个，因为2个核的最大性能和收敛效率都是最优的。

实验数据

学习率r=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/x);

第一层第二层和卷积核的权重的初始化的x分别为1000,1000,200

 

 

con(3*3)-(49)*30*2-2			cpa
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	最大值p-max
0.4991442	0.5023601	17.914573	0.503242	0.5	894.64322	178100	2.9683333	0.9054374
0.5740598	0.4261958	1272.9146	0.7974934	0.4	1554.8693	309436	5.1572667	0.9971631
0.6109585	0.3892448	1427.2714	0.9536002	0.3	1593.5427	317115	5.28525	0.9981087
0.6774132	0.3224089	1559.201	0.9703102	0.2	1652.1759	328945	5.4824167	0.9981087
0.7639885	0.2359505	1718.005	0.9784288	0.1	1729.1809	344650	5.7441667	0.9981087
0.8484488	0.1515659	2196.4422	0.9780344	0.01	1897.6834	377640	6.294	0.9971631
0.8286216	0.171378	3121.9749	0.9844185	0.001	2260.1809	449776	7.4962667	0.9985816
0.798585	0.2014137	3145.995	0.9881892	9.00E-04	2296.3819	456986	7.6164333	0.9981087
0.7635151	0.2364872	3234.5729	0.9868729	8.00E-04	1702.6734	338848	5.6474667	0.9985816
0.8187556	0.1812439	3323.3719	0.9881963	7.00E-04	2352.5075	468149	7.8024833	0.9985816
0.8188199	0.1811805	3390.7889	0.9829051	6.00E-04	2386.8995	474994	7.9165667	0.9981087
0.869077	0.1309228	3510.9196	0.9787567	5.00E-04	2404.8995	478575	7.97625	0.9981087
0.9243685	0.0756311	3657.8693	0.9836107	4.00E-04	2516.1608	500718	8.3453	0.9985816
0.9294495	0.0705505	3734.407	0.9803295	3.00E-04	2547.5276	506973	8.44955	0.9995272
0.864205	0.1357951	4094.7638	0.9759863	2.00E-04	2661.3367	529608	8.8268	0.9985816
0.7386513	0.2613488	4911.5729	0.9850553	1.00E-04	2248.9347	447540	7.459	0.9985816
0.8190485	0.1809514	4893.2412	0.9857229	9.00E-05	3025.6884	602113	10.035217	0.9995272
0.7889079	0.2110922	5162.2563	0.9856612	8.00E-05	3070.1005	610965	10.18275	0.9985816
0.8090114	0.1909886	5260.407	0.9855542	7.00E-05	3067.9598	610540	10.175667	0.9995272
0.8441875	0.1558126	5498.7236	0.9819452	6.00E-05	3160.402	628952	10.482533	0.9990544
0.9145402	0.0854598	5636.9146	0.9837557	5.00E-05	3214.0201	639590	10.659833	0.9990544
0.8441981	0.1558019	6084.8442	0.9804436	4.00E-05	2850.5276	567257	9.4542833	0.9990544
0.8190796	0.1809204	6630.5879	0.9804673	3.00E-05	3691.0151	734513	12.241883	0.9990544
0.7839104	0.2160896	7487.6985	0.9848581	2.00E-05	4037.6633	803512	13.391867	0.9990544
0.7185892	0.2814108	9588.9899	0.9863383	1.00E-05	4896.5327	974430	16.2405	1
0.7286397	0.2713602	10059.462	0.9874716	9.00E-06	4265.5427	848843	14.147383	0.9995272
0.7437155	0.2562845	10449.889	0.9874503	8.00E-06	5297.8392	1054270	17.571167	0.9990544
0.753766	0.246234	11049.116	0.9876688	7.00E-06	5531.8291	1100851	18.347517	0.9995272
0.7688416	0.2311584	11802.126	0.9870345	6.00E-06	4819.3568	959054	15.984233	0.9995272
0.7839173	0.2160827	12674.02	0.9861554	5.00E-06	6054.3719	1204870	20.081167	0.9995272
0.7336668	0.2663332	15614.482	0.9856065	4.00E-06	7127.1407	1418306	23.638433	1
0.7738679	0.2261321	17613.286	0.987733	3.00E-06	7804.392	1553079	25.88465	1
0.7939689	0.2060311	24346.503	0.9902491	2.00E-06	10443.633	2078302	34.638367	0.9995272
0.778894	0.221106	41275.352	0.9919836	1.00E-06	16988.382	3380705	56.345083	1

 

con(3*3)*2-(49*2)*30*2-2			cpa
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	最大值p-max
0.4984601	0.4988251	18.356784	0.5279756	0.5	973.60302	193749	3.22915	0.9687943
0.5466968	0.4530995	155.21106	0.8965323	0.4	1128.4372	224575	3.7429167	0.9943262
0.5951047	0.4058347	99	0.9722537	0.3	1081.8844	215342	3.5890333	0.9957447
0.6440861	0.3556425	97.246231	0.9787139	0.2	1085.6784	216050	3.6008333	0.9966903
0.6386903	0.3605568	116.82915	0.9860579	0.1	1089.2513	217136	3.6189333	0.9957447
0.9568039	0.0432011	298.80402	0.9859035	0.01	1180.7688	235050	3.9175	0.9971631
0.9138143	0.0861814	1387.6583	0.9938677	0.001	1667.1558	331779	5.52965	0.9990544
0.8486613	0.1513367	1521.3116	0.9954358	9.00E-04	1699.2211	338161	5.6360167	0.9981087
0.7433548	0.2566411	1683.397	0.9960464	8.00E-04	1773.5477	352937	5.8822833	0.9985816
0.5526884	0.4473117	1933.0452	0.9964337	7.00E-04	1922.0854	382496	6.3749333	0.9990544
0.4422747	0.5577228	2254.794	0.996581	6.00E-04	1442.2915	287016	4.7836	0.9985816
0.5527144	0.4472844	2749.6181	0.9969445	5.00E-04	2277.4774	453251	7.5541833	0.9990544
0.7435355	0.256465	3439.6583	0.9975053	4.00E-04	2603.7387	518145	8.63575	0.9995272
0.7134473	0.2865525	4653.0854	0.9977167	3.00E-04	3124.4523	621784	10.363067	0.9995272
0.5627898	0.4372103	7342.8543	0.9977476	2.00E-04	4393.4673	874301	14.571683	0.9995272
0.6833835	0.3166172	15506.523	0.9984414	1.00E-04	7331.2261	1458917	24.315283	1
0.6431918	0.3568082	13203.487	0.998477	9.00E-05	6940.2513	1381125	23.01875	1
0.738658	0.2613421	13376.839	0.9986647	8.00E-05	6964.0754	1385852	23.097533	1
0.7286127	0.271387	14084.156	0.9987479	7.00E-05	6858.8844	1364918	22.748633	1
0.7235927	0.2764073	15846.543	0.9987503	6.00E-05	8265.5678	1644864	27.4144	1
0.7939424	0.2060575	16720.432	0.9988667	5.00E-05	7847.6583	1561684	26.028067	1
0.7587739	0.241226	19631.955	0.9987455	4.00E-05	10003.613	1990720	33.178667	1
0.7889282	0.2110718	22636.035	0.9989071	3.00E-05	10208.442	2031485	33.858083	1
0.7537589	0.246241	28829.749	0.9988643	2.00E-05	13679.598	2722242	45.3707	1
0.7587891	0.241211	39033.603	0.9989498	1.00E-05	17854.095	3552968	59.216133	1
0.7989898	0.2010102	36848.085	0.9989641	9.00E-06	17032.834	3389535	56.49225	1
0.7236146	0.2763853	37868.141	0.9990401	8.00E-06	16425.859	3268748	54.479133	1
0.6884396	0.3115604	37583.121	0.9989498	7.00E-06	15341.171	3052901	50.881683	1
0.6180891	0.381911	39620.457	0.9988833	6.00E-06	16804.392	3344075	55.734583	1
0.6532648	0.3467352	43121.608	0.998957	5.00E-06	17687.578	3519843	58.66405	1
0.6532651	0.3467349	44738.422	0.9989641	4.00E-06	17258.553	3434452	57.240867	1
0.6683407	0.3316593	49216.608	0.9989261	3.00E-06	19840.442	3948248	65.804133	1
0.6130649	0.3869351	56637.367	0.998938	2.00E-06	23083.397	4593599	76.559983	1
0.6633163	0.3366837	70958.894	0.9989926	1.00E-06	28386.724	5648974	94.149567	1

 

con(3*3)*3-(49*3)*30*2-2			cpa
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	最大值p-max
0.4992652	0.4997364	16.407035	0.5106597	0.5	927.36683	184549	3.0758167	0.9579196
0.5884402	0.4128762	135.27638	0.8834884	0.4	906.9799	180489	3.00815	0.9981087
0.6537869	0.3473519	84.81407	0.9794124	0.3	884.57789	176049	2.93415	0.9971631
0.7126646	0.2878306	92.437186	0.9882129	0.2	886.78392	176471	2.9411833	0.9976359
0.7324443	0.2669509	116.30151	0.9907267	0.1	898.50754	178805	2.9800833	0.9971631
0.9666153	0.0334501	326.79397	0.9908051	0.01	996.46231	198297	3.30495	0.9981087
0.9639662	0.0360329	1500.9146	0.9944712	0.001	1523.9246	303262	5.0543667	0.9990544
0.8135525	0.1864475	1674.1809	0.9955712	9.00E-04	1604.3166	319259	5.3209833	0.9990544
0.6078763	0.3921203	1929.9246	0.9959229	8.00E-04	1717.3216	341763	5.69605	0.9990544
0.4774135	0.5225865	2141.201	0.9960678	7.00E-04	1814.6935	361125	6.01875	0.9990544
0.4573306	0.5426685	2411.7437	0.9965382	6.00E-04	2031.603	404291	6.7381833	0.9990544
0.6832444	0.3167547	2905.4774	0.9970895	5.00E-04	2393.2161	476252	7.9375333	0.9995272
0.8037951	0.1962048	3576.0704	0.9975456	4.00E-04	2614.5628	520315	8.6719167	1
0.7435807	0.2564185	4458.1709	0.9977714	3.00E-04	3001.8191	597362	9.9560333	1
0.7436279	0.2563727	6998.9347	0.9974791	2.00E-04	4155.4372	826940	13.782333	1
0.668311	0.3316889	14384.759	0.9978403	1.00E-04	6296.3819	1252987	20.883117	1
0.6984594	0.3015405	14424.467	0.9984865	9.00E-05	7348.8342	1462419	24.37365	1
0.6783655	0.3216347	14449.583	0.9982085	8.00E-05	6849.8291	1363116	22.7186	1
0.6030017	0.3969984	15548.824	0.9984105	7.00E-05	7951.5176	1582352	26.372533	1
0.713544	0.2864561	17738.523	0.9987431	6.00E-05	8147.9397	1621440	27.024	1
0.6532518	0.3467483	19301.136	0.9987241	5.00E-05	9678.6985	1926078	32.1013	1
0.768824	0.231176	20924.327	0.9989617	4.00E-05	9483.7588	1887268	31.454467	1
0.6331581	0.3668419	23709.357	0.9988144	3.00E-05	11334.422	2255582	37.593033	1
0.6984847	0.301515	31211.613	0.9989736	2.00E-05	14235.92	2832948	47.2158	1
0.6432133	0.3567867	43837.337	0.9989142	1.00E-05	20381.035	4055845	67.597417	1
0.6582887	0.3417113	42105.221	0.9987954	9.00E-06	20147.849	4009424	66.823733	1
0.457287	0.5427129	51632.447	0.9963648	8.00E-06	22352.935	4448237	74.137283	1
0.6582894	0.3417106	49078.015	0.9985032	7.00E-06	21170.518	4212936	70.2156	1
0.7135654	0.2864345	47897.337	0.9985222	6.00E-06	20763.447	4131945	68.86575	1
0.6532649	0.3467351	49984.935	0.9985032	5.00E-06	21364.93	4251652	70.860867	1
0.6080394	0.3919606	53647.573	0.998553	4.00E-06	22341.698	4446000	74.1	1
0.6381902	0.3618098	57838.588	0.9987669	3.00E-06	24684.849	4912301	81.871683	1
0.5175879	0.4824121	67256.789	0.9988049	2.00E-06	28073.603	5586663	93.11105	1
0.6130651	0.3869349	85642.864	0.9987312	1.00E-06	35519.231	7068329	117.80548	1

 

con(3*3)*4-(49*4)*30*2-2			cpa
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	最大值p-max
0.5002834	0.4981237	17.482412	0.5228435	0.5	926.40704	184362	3.0727	0.9546099
0.5892083	0.4109681	124.34171	0.8674745	0.4	1020.9648	203175	3.38625	0.9976359
0.6582225	0.3419309	86.361809	0.9801205	0.3	1039.1106	206793	3.44655	0.9981087
0.7159889	0.2840174	85.959799	0.989111	0.2	989.50251	196921	3.2820167	0.9981087
0.7125465	0.2872574	113.48241	0.9918505	0.1	992.77889	197590	3.2931667	0.9976359
0.9911855	0.0088149	338.24623	0.9919075	0.01	1107.9749	220487	3.6747833	0.9981087
0.8636796	0.1363193	1359.9397	0.9857206	0.001	1513.8241	301271	5.0211833	0.9985816
0.678101	0.3219019	1741.0402	0.9923827	9.00E-04	1708.7136	340040	5.6673333	0.9985816
0.4774182	0.5225791	2128.1256	0.9945591	8.00E-04	1893.3668	376797	6.27995	0.9990544
0.4623561	0.5376444	2405.0151	0.9954382	7.00E-04	1997.0402	397415	6.6235833	0.9995272
0.6279977	0.3720034	2881.5528	0.9965668	6.00E-04	2289.196	455551	7.5925167	0.9995272
0.6782241	0.3217751	3194.9397	0.9969968	5.00E-04	2414.9497	480575	8.0095833	0.9995272
0.8288998	0.1710996	4063.3116	0.9975361	4.00E-04	2823.5226	561897	9.36495	0.9995272
0.7134475	0.2865519	5366.3015	0.9972795	3.00E-04	2615.9347	520572	8.6762	1
0.7134873	0.2865119	7857.2864	0.9966879	2.00E-04	4671.3568	929601	15.49335	1
0.633141	0.3668592	15661.156	0.9976027	1.00E-04	8476.7538	1686877	28.114617	1
0.6532407	0.3467603	15587.477	0.9981824	9.00E-05	7736.8543	1539652	25.660867	1
0.6783646	0.321635	15865.03	0.9984438	8.00E-05	8603.005	1712000	28.533333	1
0.6130498	0.3869503	17521.437	0.9985079	7.00E-05	8540.804	1699623	28.32705	1
0.6532492	0.3467511	18500.673	0.9987051	6.00E-05	9851.9246	1960536	32.6756	1
0.6733504	0.3266496	19821.92	0.9986006	5.00E-05	9446.6281	1879888	31.331467	1
0.6130567	0.3869432	22956	0.9986338	4.00E-05	11203.487	2229511	37.158517	1
0.5879345	0.4120654	27418.879	0.9986932	3.00E-05	14380.573	2861735	47.695583	1
0.5979862	0.402014	34490.317	0.9987146	2.00E-05	16701.608	3323622	55.3937	1
0.603013	0.396987	48120.171	0.9986837	1.00E-05	22269.231	4431594	73.8599	1
0.5527628	0.4472371	46714.543	0.9986481	9.00E-06	22824.241	4542025	75.700417	1
0.6281387	0.3718613	47212.638	0.9988001	8.00E-06	22742.538	4525765	75.429417	1
0.6030137	0.3969863	49336.06	0.9987479	7.00E-06	23251.171	4626983	77.116383	1
0.5477381	0.452262	52680.538	0.9987289	6.00E-06	24958.196	4966681	82.778017	1
0.5879388	0.4120611	52986.186	0.9988144	5.00E-06	26070.332	5187996	86.4666	1
0.6633154	0.3366846	58730.332	0.9987289	4.00E-06	27308.347	5434372	90.572867	1
0.5226129	0.4773871	80131.437	0.9976146	3.00E-06	38146.819	7591235	126.52058	1
0.5326632	0.4673368	83585.859	0.9980921	2.00E-06	39802.779	7920757	132.01262	1
0.5175879	0.4824121	101804.81	0.9984485	1.00E-06	46908.764	9334844	155.58073	1

 

con(3*3)*5-(49*5)*30*2-2			cpa
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	最大值p-max
0.502419	0.4987094	15.547739	0.5252337	0.5	811.94472	161592	2.6932	0.9347518
0.593106	0.4064401	117.74372	0.830127	0.4	860.52261	171244	2.8540667	0.9962175
0.6662483	0.3337374	85.241206	0.9772218	0.3	842.88945	167735	2.7955833	0.9976359
0.7607087	0.2384696	87.366834	0.988346	0.2	844.9397	168143	2.8023833	0.9985816
0.758232	0.2418952	115.65829	0.9929838	0.1	861.19598	171378	2.8563	0.9976359
0.9813115	0.0186904	338.8392	0.9912541	0.01	981.41206	195317	3.2552833	0.9985816
0.9740381	0.0259685	1463.8492	0.9875144	0.001	1570.8894	312623	5.2103833	0.9985816
0.7985059	0.2014929	1752.8995	0.9928199	9.00E-04	1724.7387	343223	5.7203833	0.9990544
0.6229303	0.3770704	2020.1005	0.9949749	8.00E-04	1869.9347	372117	6.20195	0.9990544
0.5928377	0.4071632	2250.4523	0.995652	7.00E-04	1989.5779	395942	6.5990333	0.9990544
0.6681521	0.3318476	2584.4573	0.9965501	6.00E-04	2168.4322	431518	7.1919667	0.9995272
0.7836555	0.2163438	3009.8894	0.9969659	5.00E-04	2390.9296	475811	7.9301833	0.9995272
0.8640518	0.1359472	3557.0352	0.9974886	4.00E-04	2677.603	532843	8.8807167	0.9995272
0.6782894	0.3217107	4798.6784	0.9968875	3.00E-04	3349.8442	666635	11.110583	0.9995272
0.7637195	0.2362798	6804.7437	0.9959823	2.00E-04	4394.2362	874469	14.574483	1
0.5929474	0.4070529	12383.598	0.9964884	1.00E-04	6846.1357	1362381	22.70635	1
0.5577792	0.4422214	14052.729	0.9969374	9.00E-05	8173.1759	1626494	27.108233	1
0.5929508	0.4070492	15632.03	0.9974839	8.00E-05	8996.2111	1790246	29.837433	1
0.6230988	0.376901	17895.95	0.9979709	7.00E-05	10741.392	2137537	35.625617	1
0.653249	0.3467511	19363.005	0.998439	6.00E-05	10036.598	1997289	33.28815	1
0.5728566	0.4271433	21446.03	0.9983891	5.00E-05	11969.06	2381851	39.697517	1
0.6683287	0.3316711	31802.839	0.9955546	4.00E-05	17303.583	3443431	57.390517	1
0.537686	0.4623137	37657.161	0.9980969	3.00E-05	19683.015	3916920	65.282	1
0.5075373	0.4924627	39645.352	0.9982561	2.00E-05	21157.714	4210403	70.173383	1
0.452262	0.5477379	53363.085	0.9980612	1.00E-05	29097.874	5790508	96.508467	1
0.5075375	0.4924625	51339.673	0.9984509	9.00E-06	26888.935	5350899	89.18165	1
0.4924624	0.5075377	51188.322	0.9983962	8.00E-06	27097.266	5392356	89.8726	1
0.457287	0.542713	77134.93	0.9974744	7.00E-06	40812.357	8121660	135.361	1
0.5678385	0.4321615	68720.065	0.9979353	6.00E-06	35771.01	7118433	118.64055	1
0.5628135	0.4371865	68404.271	0.9984461	5.00E-06	36759.96	7315232	121.92053	1
0.5778889	0.4221111	70417.965	0.9983273	4.00E-06	42360.342	8429710	140.49517	1
0.4623118	0.5376882	79051.045	0.9986766	3.00E-06	42580.719	8473579	141.22632	1
0.5125628	0.4874372	87650.271	0.9987027	2.00E-06	48035.603	9559094	159.31823	1
0.8442205	0.1557795	117332.54	0.9969303	1.00E-06	62024.191	12342816	205.7136	1

 

 

con(3*3)*6-(49*6)*30*2-2			cpa
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	最大值p-max
0.501056	0.4970626	17.467337	0.5275835	0.5	870.58291	173264	2.8877333	0.9867612
0.5874694	0.4137517	113.80905	0.8256673	0.4	926.08543	184291	3.0715167	0.9962175
0.6751543	0.3240751	81.59799	0.9764377	0.3	907.21106	180535	3.0089167	0.9981087
0.7231123	0.2766511	88.311558	0.9892797	0.2	906.46734	180389	3.0064833	0.9976359
0.776154	0.2235614	117.47739	0.9927676	0.1	923.88442	183853	3.0642167	0.9976359
0.9862619	0.0137143	339.51759	0.9875833	0.01	1063.7889	211694	3.5282333	0.9990544
0.9539836	0.0460196	1509.4573	0.9798805	0.001	1773.4322	352913	5.8818833	0.9990544
0.883782	0.1162176	1745.9196	0.9906174	9.00E-04	1916.4975	381399	6.35665	0.9995272
0.6379837	0.362016	1998.7387	0.9936871	8.00E-04	1030.1106	204992	3.4165333	0.9990544
0.5828051	0.4171965	2339.8693	0.9952196	7.00E-04	2231.9849	444166	7.4027667	0.9995272
0.6380373	0.3619622	2677.9648	0.9961557	6.00E-04	2442.4774	486054	8.1009	0.9995272
0.7937006	0.2063007	3087.598	0.9968946	5.00E-04	2692.9045	535904	8.9317333	0.9995272
0.8439627	0.1560347	3701.4874	0.996859	4.00E-04	3062.4221	609423	10.15705	0.9995272
0.6782909	0.3217102	5027.9849	0.9966333	3.00E-04	3790.2362	754258	12.570967	1
0.7185103	0.2814898	7065.5126	0.9947088	2.00E-04	4202.1709	836236	13.937267	1
0.5175832	0.4824166	13033.332	0.9957043	1.00E-04	8218.0653	1635397	27.256617	1
0.5376808	0.4623192	13926.789	0.9971251	9.00E-05	8993.4372	1789701	29.82835	1
0.643194	0.3568054	15787.98	0.9974934	8.00E-05	10204.588	2030715	33.84525	1
0.5829033	0.4170966	17362.663	0.9977975	7.00E-05	10422.186	2074016	34.566933	1
0.5577818	0.4422181	20086.668	0.9979923	6.00E-05	11844.819	2357120	39.285333	1
0.5477334	0.4522664	23328.317	0.998161	5.00E-05	14664.211	2918179	48.636317	1
0.4974872	0.5025127	26592.92	0.9982774	4.00E-05	15998.704	3183743	53.062383	1
0.4924626	0.5075377	31323.106	0.9982632	3.00E-05	18347.794	3651211	60.853517	1
0.5125622	0.4874379	41799.608	0.99842	2.00E-05	24354.166	4846479	80.77465	1
0.4824123	0.5175876	60223.447	0.9983487	1.00E-05	35398.518	7044307	117.40512	1
0.4271368	0.5728632	60674.99	0.998363	9.00E-06	34985.834	6962184	116.0364	1
0.7035147	0.2964853	78991.548	0.9962721	8.00E-06	46034.452	9160861	152.68102	1
0.7135651	0.2864348	62342.04	0.9980446	7.00E-06	35925.714	7149220	119.15367	1
0.6582897	0.3417102	63644.176	0.9980541	6.00E-06	36644.442	7292245	121.53742	1
0.6080392	0.3919608	65834.437	0.9983749	5.00E-06	38265.729	7614881	126.91468	1
0.6532652	0.3467348	75075.065	0.9987004	4.00E-06	43278.563	8612443	143.54072	1
0.5628137	0.4371863	82619.769	0.99866	3.00E-06	47495.97	9451716	157.5286	1
0.6231151	0.3768849	91820.995	0.99866	2.00E-06	52878.578	10522837	175.38062	1
0.7487433	0.2512567	131001.35	0.9967545	1.00E-06	76600.392	15243526	254.05877	1

con(3*3)*7-(49*7)*30*2-2			cpa
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	最大值p-max
0.4971291	0.5010397	12.60804	0.5127529	0.5	988.92965	196812	3.2802	0.9706856
0.5996055	0.4001422	110.52261	0.7895553	0.4	1057.2261	210388	3.5064667	0.9966903
0.6883169	0.3111753	84.592965	0.9727384	0.3	1033.4724	205662	3.4277	0.9981087
0.7744565	0.2256916	85.276382	0.9897692	0.2	1032.3618	205441	3.4240167	0.9985816
0.7646914	0.2350993	124.9196	0.9933877	0.1	1049.0503	208776	3.4796	0.9981087
0.9863947	0.0136193	350.22111	0.9876451	0.01	1201.9799	239194	3.9865667	0.9990544
0.9690574	0.0309428	1473.794	0.9729427	0.001	1221.8643	243151	4.0525167	0.9990544
0.8286096	0.171394	1758.8995	0.9851456	9.00E-04	2152.0704	428262	7.1377	0.9990544
0.6128992	0.3871023	2082.3819	0.9924754	8.00E-04	2365.4623	470728	7.8454667	0.9990544
0.5125463	0.487452	2384.5025	0.9944926	7.00E-04	2745.1005	546279	9.10465	0.9995272
0.6882341	0.3117671	2801.3116	0.9956924	6.00E-04	3178.7839	632582	10.543033	0.9995272
0.7083501	0.2916508	3136.9698	0.9955831	5.00E-04	3414.5025	679489	11.324817	0.9995272
0.8037931	0.1962065	3743.7688	0.9967568	4.00E-04	3764.0653	749051	12.484183	0.9995272
0.5878849	0.412115	5106.8995	0.9953431	3.00E-04	4109.5075	817793	13.629883	1
0.7134868	0.286512	7001.5779	0.993775	2.00E-04	5978.7889	1189787	19.829783	1
0.4723641	0.5276354	12325.96	0.9951008	1.00E-04	9786.8291	1947581	32.459683	1
0.5376815	0.4623187	13759.533	0.9963102	9.00E-05	11527.457	2293981	38.233017	1
0.5577795	0.4422207	14914.91	0.9966309	8.00E-05	10936.015	2176285	36.271417	1
0.4874378	0.5125625	17504.095	0.9975076	7.00E-05	13092.317	2605371	43.42285	1
0.6030036	0.3969966	19431.678	0.9976526	6.00E-05	14356.563	2856957	47.61595	1
0.5376843	0.4623156	21723.905	0.9980517	5.00E-05	16257.889	3235320	53.922	1
0.4974871	0.5025127	25417.492	0.9980256	4.00E-05	18494.161	3680339	61.338983	1
0.4874375	0.5125625	31166.216	0.9982299	3.00E-05	22442.769	4466111	74.435183	1
0.5376871	0.4623129	39314.266	0.9982513	2.00E-05	28302.357	5632170	93.8695	1
0.8542649	0.1457352	73658.322	0.99866	1.00E-05	46679.623	9289248	154.8208	1
0.6984892	0.3015108	62954.739	0.9984889	9.00E-06	39966.281	7953292	132.55487	1
0.5979885	0.4020115	64876.879	0.9986315	8.00E-06	45212.804	8997348	149.9558	1
0.5678383	0.4321616	69872.558	0.9985602	7.00E-06	48378.196	9627261	160.45435	1
0.6281393	0.3718607	77103.864	0.9986837	6.00E-06	52354.628	10418572	173.64287	1
0.5376881	0.4623119	75845.513	0.9986124	5.00E-06	51846.879	10317530	171.95883	1
0.8894444	0.1105556	88052.658	0.9957756	4.00E-06	55477.472	11040019	184.00032	1
0.6934663	0.3065337	91937.769	0.9975361	3.00E-06	57538.437	11450151	190.83585	1
0.6381904	0.3618095	97789.352	0.9980493	2.00E-06	67650.095	13462370	224.37283	1
0.6130651	0.3869349	129437.56	0.9982988	1.00E-06	89419.121	17794421	296.57368	1

 

con(3*3)*8-(49*8)*30*2-2			cpa
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	最大值p-max
0.5001463	0.4975267	14.236181	0.5360657	0.5	1174.1357	233662	3.8943667	0.9820331
0.601897	0.3990353	106.59799	0.7497986	0.4	1274.6834	253668	4.2278	0.9981087
0.6882475	0.3107847	79.839196	0.9708923	0.3	1230.9447	244964	4.0827333	0.9981087
0.7821435	0.218235	80.633166	0.9887166	0.2	1293.0603	257324	4.2887333	0.9981087
0.7848443	0.2149465	117.65829	0.9934923	0.1	1343.8442	267444	4.4574	0.9981087
0.9813175	0.0186511	347.95477	0.9869275	0.01	1506.6935	299838	4.9973	0.9995272
0.9740762	0.0259265	1451.7889	0.9615857	0.001	2354.8945	468628	7.8104667	0.9990544
0.7884817	0.2115207	1784.8291	0.9808736	9.00E-04	2526.0402	502685	8.3780833	0.9990544
0.6179132	0.3820865	2139.2412	0.9904677	8.00E-04	2790.5126	555313	9.2552167	0.9995272
0.537636	0.462363	2390.9799	0.993181	7.00E-04	2996.9849	596402	9.9400333	0.9995272
0.597879	0.402118	2781.1608	0.9949012	6.00E-04	3332.7739	663225	11.05375	0.9995272
0.7334545	0.2665463	3252.9045	0.9960298	5.00E-04	2720.4472	541378	9.0229667	0.9995272
0.7987748	0.2012279	3885.201	0.9959609	4.00E-04	4161.5729	828165	13.80275	0.9995272
0.7486054	0.2513941	3931.8844	0.9648479	3.00E-04	3632.8392	722952	12.0492	0.9995272
0.4974828	0.5025161	9039.8241	0.9880775	2.00E-04	7026.2111	1398233	23.303883	1
0.7085037	0.2914961	18030.437	0.9960345	1.00E-04	13143.327	2615537	43.592283	1
0.6180692	0.3819299	19101.362	0.9965929	9.00E-05	13315.437	2649773	44.162883	1
0.5979739	0.4020261	19248.337	0.9974839	8.00E-05	14002.246	2786447	46.440783	1
0.6130506	0.3869502	21427.548	0.9975932	7.00E-05	15327.266	3050142	50.8357	1
0.5778799	0.4221196	23328.92	0.998047	6.00E-05	16181.271	3220073	53.667883	1
0.5276353	0.4723644	25601.452	0.9978973	5.00E-05	18228.638	3627514	60.458567	1
0.4974876	0.5025126	29984.538	0.9980826	4.00E-05	20917.975	4162677	69.37795	1
0.4974873	0.5025126	33767.291	0.9983345	3.00E-05	23035.141	4584011	76.400183	1
0.4723627	0.5276373	43519.91	0.9983558	2.00E-05	30702.08	6109714	101.82857	1
0.4472371	0.5527629	63562.337	0.9985317	1.00E-05	43157.121	8588283	143.13805	1
0.4020116	0.5979884	65050.985	0.9983036	9.00E-06	44379.211	8831480	147.19133	1
0.8643163	0.1356837	77089.065	0.996745	8.00E-06	52901.477	10527396	175.4566	1
0.7537657	0.2462343	66253.085	0.9974435	7.00E-06	45749.377	9104126	151.73543	1
0.7035154	0.2964845	68074	0.998382	6.00E-06	47599.975	9472411	157.87352	1
0.6633151	0.3366849	75117.256	0.9980208	5.00E-06	55087.513	10962417	182.70695	1
0.6633154	0.3366846	76259.136	0.9982727	4.00E-06	58492.899	11640088	194.00147	1
0.8643196	0.1356804	97506.925	0.996859	3.00E-06	67885.874	13509290	225.15483	1
0.7236173	0.2763827	117858.88	0.9970348	2.00E-06	86917.412	17296565	288.27608	1
0.6633163	0.3366837	143227.49	0.997586	1.00E-06	107430.87	21378744	356.3124	1

 

con(3*3)*9-(49*9)*30*2-2			cpa
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	最大值p-max
0.5004345	0.4969997	14.623116	0.5396533	0.5	961.28141	191296	3.1882667	0.9910165
0.5935062	0.4070302	106.54271	0.7164261	0.4	1067.0402	212345	3.5390833	0.9966903
0.7005314	0.3000457	80.693467	0.9678321	0.3	1053.598	209666	3.4944333	0.9981087
0.7944988	0.2049604	85.889447	0.9880086	0.2	1049.7789	208923	3.48205	0.9976359
0.7428472	0.2569587	117.50754	0.9935612	0.1	1078.7085	214666	3.5777667	0.9985816
0.9912877	0.0086961	351.40201	0.986153	0.01	1257.4573	250234	4.1705667	0.9985816
0.9841139	0.0158891	1397.1156	0.9507158	0.001	2077.8945	413535	6.89225	0.9985816
0.9038809	0.0961187	1689.8291	0.9716859	9.00E-04	2315.3266	460751	7.6791833	0.9990544
0.6530397	0.3469581	2169.9799	0.9890968	8.00E-04	2579.3015	513298	8.5549667	0.9990544
0.5326165	0.4673838	2421.4372	0.9939485	7.00E-04	2751.1859	547486	9.1247667	0.9990544
0.5928609	0.4071387	2799.4774	0.9934329	6.00E-04	2017.1407	401411	6.6901833	0.9995272
0.78868	0.2113211	3256.0452	0.9952885	5.00E-04	3402.8995	677178	11.2863	0.9995272
0.7686425	0.2313571	3781.995	0.9944545	4.00E-04	3756.5176	747547	12.459117	0.9995272
0.6381136	0.3618876	5129.3216	0.9940578	3.00E-04	4757.3668	946733	15.778883	1
0.5929282	0.4070717	7125.9598	0.9921261	2.00E-04	5739.3869	1142154	19.0359	1
0.4723621	0.5276377	12689.749	0.9955356	1.00E-04	10491.95	2087901	34.79835	1
0.5025093	0.4974911	13822.01	0.9952671	9.00E-05	10480.879	2085712	34.761867	1
0.4874366	0.5125632	15260.698	0.9960916	8.00E-05	11756.834	2339611	38.993517	1
0.5125593	0.4874406	17085.739	0.9964456	7.00E-05	13843.829	2754923	45.915383	1
0.5728555	0.4271448	18479.136	0.9972415	6.00E-05	14083.709	2802658	46.710967	1
0.5326597	0.4673405	22090.452	0.9973271	5.00E-05	16906.704	3364436	56.073933	1
0.5628089	0.4371908	25991.136	0.9975789	4.00E-05	19535.799	3887630	64.793833	1
0.5427109	0.4572892	32577.467	0.9979804	3.00E-05	24581.161	4891667	81.527783	1
0.4824124	0.5175875	39049.779	0.9982988	2.00E-05	29246.05	5819980	96.999667	1
0.4874373	0.5125627	62636.548	0.9982062	1.00E-05	49261.045	9802949	163.38248	1
0.8040151	0.1959848	86731.849	0.9974554	9.00E-06	66086.025	13151122	219.18537	1
0.5527631	0.4472369	90758.111	0.9980232	8.00E-06	68435.97	13618759	226.97932	1
0.532663	0.467337	79820.975	0.9979709	7.00E-06	60225.889	11984961	199.74935	1
0.5427131	0.4572869	75305.739	0.9981872	6.00E-06	57658.015	11473948	191.23247	1
0.8693433	0.1306567	85185.03	0.9982537	5.00E-06	63082.186	12553371	209.22285	1
0.6733656	0.3266344	99206.302	0.9982584	4.00E-06	76016.95	15127389	252.12315	1
0.6633157	0.3366843	89498.804	0.9983202	3.00E-06	72579.111	14443275	240.72125	1
0.8341696	0.1658304	125675.34	0.9965858	2.00E-06	100546.08	20008681	333.47802	1
0.7236177	0.2763823	145569.59	0.9980113	1.00E-06	124728.04	24820880	413.68133	1