欢迎任何改进意见或提出全新方法! 谢谢!

牛能编程我至少还要两年后才有这水平

这里应该有不少Eviews的大牛们呀! 不要光看不回贴呀! 任何想法和建议都是好的! 在此谢过!

受EViews内存空间的约束，!t增大，系统会到硬盘暂存数据，必然很慢。

楼主的程序很节省内存。下面是学习楼主程序的习作。用3个循环生成待评价的向量（序列）。

!t=10 !track = 0.90 create abc u 1 !t series markt = 2*@round(rnd) - 1 stom(markt,vecmarkt) scalar pb=2*!track*!t - !t !Q=3^(!t) matrix(!Q,!t) mm matrix(!t,!Q) m group success for !j=1 to !t for !i=1 to 3^(!j) for !k=1+(!i-1)*3^(!t-!j) to !i*3^(!t-!j) if @mod(!i,3)=1 then mm(!k,!j)=-1 endif if @mod(!i,3)=2 then mm(!k,!j)=0 endif if @mod(!i,3)=0 then mm(!k,!j)=1 endif next next next m=@transpose(mm) !k=0 for !j=1 to !Q vector v1=@columnextract(m, !j) scalar ergeb = @inner(v1,vecmarkt) if ergeb>=pb then !k=!k+1 series s!k mtos(v1,s{!k}) success.add s{!k} endif next freeze(strattable) success.table

在EViews4.1和EViews5.1（企业版）下都通过。

以下是对楼主程序的少许“改进”，在EViews5.1下通过。

'设定工作文件的区间长度，!t=10不到一分钟，!t=14约14分钟，!t增加1决不止增加一分钟 !t=14 !track = 0.90 create abc u 1 !t '为工作变量赋初始值 series markt = 2*@round(rnd) - 1 stom(markt,vecmarkt) scalar pb=2*!track*!t - !t group success rowvector(!t) strat !i=0 '设定循环 for !v1 = -1 to 1 step 1 for !v2 = -1 to 1 step 1 for !v3 = -1 to 1 step 1 for !v4 = -1 to 1 step 1 for !v5 = -1 to 1 step 1 for !v6 = -1 to 1 step 1 for !v7 = -1 to 1 step 1 for !v8 = -1 to 1 step 1 for !v9 = -1 to 1 step 1 for !v10 = -1 to 1 step 1 for !v11 = -1 to 1 step 1 for !v12 = -1 to 1 step 1 for !v13 = -1 to 1 step 1 for !v14 = -1 to 1 step 1 'for !v15 = -1 to 1 step 1 'for !v16 = -1 to 1 step 1 'for !v17 = -1 to 1 step 1 'for !v18 = -1 to 1 step 1 'for !v19 = -1 to 1 step 1 'for !v20 = -1 to 1 step 1

'循环体开始 for !pos = 1 to !t strat(!pos) = !v{!pos} next scalar ergeb = @inner(strat,vecmarkt)

if ergeb >= pb then !i = !i + 1 series s!i mtos(strat, s!i) success.add s!i endif '循环体结束

'next '20 'next '19 'next '18 'next '17 'next '16 'next '15 next '14 next '13 next '12 next '11 next '10 next '9 next '8 next '7 next '6 next '5 next '4 next '3 next '2 next '1 '展示处理结果 freeze(strattable) success.table

'save abc.wf1