ls b2 ar(1) ma(1)//确定了arma模型的阶数，对原来样本序列进行建模，进行预报
C公司
genr c33=d(c3)//对C公司数据先求一介差分，以使该序列成为平稳序列
scalar c_mean=@mean(c33)//求出C公司差分后序列的平均值，为后面建立arma模型准备
genr c_zero=c33-c_mean//对其进行零均值化
ls c_zero ar(1) ma(1)//对其进行arma建模，通过后面的Q检验，参差图等等检验手段确定arma(1,1)模型最为合适
expand 1980 2014//扩展序列范围，为了后面的预测作准备
smpl 1980 2003//设置样本的范围
ls c3 ar(1) ma(1)//确定了arma模型的阶数，对原来样本序列进行建模，进行预报
D公司
genr d44=d(d4)
scalar d_mean=@mean(d44)
genr d_zero=d44-d_mean
ls d4 ar(1) /对其进行arma建模，通过后面的Q检验，参差图等等检验手段确定ar(1)模型最为合适
expand 1980 2014
smpl 1980 2003
ls d4 ar(1)
Lingo代码：
MODEL:
SETS:
PRE_YEAR/1..11/:A,B,C,D;!四个公司的变量
ACT_YEAR(PRE_YEAR):A1,B2,C3,D4;!四个公司待求的变量
ENDSETS
DATA:
A=9.05192, 9.0856, 9.119406, 9.15338, 9.187396, 9.22158, 9.255892, 9.290331, 9.324899, 9.359595, 9.394421;
B=7.382793, 7.328626, 7.274856, 7.221481, 7.168498 ,7.115903, 7.063694, 7.011868, 6.960423, 6.909355, 6.858661;
C=10.82064, 10.95653, 11.09412, 11.23344, 11.37451, 11.51736, 11.66199, 11.80844, 11.95674, 12.10689, 12.25893;
D=8.828845, 8.765304, 8.702221, 8.639591, 8.577413, 8.515681, 8.454395, 8.393549, 8.333141, 8.273168, 8.213626;
ENDDATA
max=@sum(PRE_YEAR(J):A1/A+B2/B+C3/C+D4/D);!线性目标
!下面为约束条件，35.44为用概率决策定出的总公司投入的医疗费用总和
@for(PRE_YEAR(J):@sum(A(J)+B(J)+C(J)+D(J)=35.44));
@for(PRE_YEAR(J):A1(J)<=A(J)#AND#B2(J)<=B(J)#AND#C3(J)<=C(J)#AND#D4(J)<=D(J));
@for(PRE_YEAR(J):A1(J)>=0#AND#B2(J)>=0#AND#C3(J)>=0#AND#D4(J)>=0);
END
