如何在prolog列表中存储递归函数的值?

时间:2015-11-13 17:17:49

标签: list recursion prolog

我的任务是:

  

给定函数f的以下定义,创建一个Prolog程序来计算所有0<我< 32。

     
      
  • f(0)= 0
  •   
  • f(1)= 1
  •   
  • f(n)= f(n-2)+ 2 * f(n-1),n> 1。 1
  •   

到目前为止我的代码是:

problemThree(0, 0).
problemThree(1, 1).
problemThree(N, NF) :-
    N > 1,
    A is N - 2,
    B is N - 1,
    problemThree(A, AF),
    problemThree(B, BF),
    NF is AF + 2*BF.

它正在发挥作用,但显示N>的值是永远的。 20。

请告诉我如何将值存储在列表中以使程序更快。

4 个答案:

答案 0 :(得分:4)

这是一种DCG方法,它将序列生成为列表:

prob3(1, F0, F1) --> [F0, F1].
prob3(N, F0, F1) --> {N > 1, F2 is 2*F1 + F0, N1 is N-1}, [F0], prob3(N1, F1, F2).

prob3(0, [0]).
prob3(N, FS) :-
    phrase(prob3(N, 0, 1), FS).

?- prob3(10, L).
L = [0, 1, 2, 5, 12, 29, 70, 169, 408, 985] ;
false.

?- prob3(169, L).
L = [1, 2, 5, 12, 29, 70, 169, 408, 985, 2378, 5741, 13860, 33461, 80782, 195025,
..., 17280083176824678419775054525017769508908307965108250063833395641] ;
false

?- time((prob3(1000, L),false)).
% 3,011 inferences, 0.005 CPU in 0.005 seconds (100% CPU, 628956 Lips)
false.

<小时/> 请注意,对于长列表答案,SWI Prolog将缩写输出,例如:

?- prob3(20, L).
L = [0, 1, 2, 5, 12, 29, 70, 169, 408|...]

这只是SWI Prolog的一种方式,它不会使大量输出混乱。在这里,您可以使用w进行回复,它会给出完整的结果:

?- prob3(20, L).
L = [0, 1, 2, 5, 12, 29, 70, 169, 408|...] [write]   % PRESSED 'w' here
L = [0, 1, 2, 5, 12, 29, 70, 169, 408, 985, 2378, 5741, 13860, 33461, 80782, 195025, 470832, 1136689, 2744210, 6625109, 15994428] ;
false

?-

请参阅Help: I want the whole answer

答案 1 :(得分:2)

无需存储超过前两个数字!

这是我从臀部快速而肮脏的镜头:

p3(N,F) :- 
   (  N =:= 0 -> F = 0
   ;  N =:= 1 -> F = 1
   ;  N  >  1 -> N0 is N-2, p3_(N0,0,1,F)
   ).

p3_(N,F0,F1,F) :- 
   F2 is F0 + 2*F1,
   (  N =:= 0
   -> F2 = F
   ;  N0 is N-1,
      p3_(N0,F1,F2,F)
   ).

示例查询:

?- between(25,35,N), p3(N,F).
  N = 25, F = 1311738121
; N = 26, F = 3166815962
; N = 27, F = 7645370045
; N = 28, F = 18457556052
; N = 29, F = 44560482149
; N = 30, F = 107578520350
; N = 31, F = 259717522849
; N = 32, F = 627013566048
; N = 33, F = 1513744654945
; N = 34, F = 3654502875938
; N = 35, F = 8822750406821.

某些东西有点更大:

?- p3(111,F).
F = 1087817594842494380941469835430214208491185.

?- p3(123,F).
F = 42644625325266431622582204734101084193553730205.

?- p3(169,F).
F = 17280083176824678419775054525017769508908307965108250063833395641.

足够快?

?- time((between(0,1000,N), p3(N,_), false)).
% 2,006,005 inferences, 0.265 CPU in 0.265 seconds (100% CPU, 7570157 Lips)
false.

答案 2 :(得分:2)

虽然它比其他答案慢得多,但我喜欢这个具有功能精神的人:

:- use_module(library(lambda)).

f(N, FN) :-
    cont_f(N, _, FN, \_^Y^_^U^(U = Y)).

cont_f(N, FN1, FN, Pred) :-
    (   N < 2 ->
        call(Pred, 0, 1, FN1, FN)
    ;
        N1 is N - 1,
        P = \X^Y^Y^U^(U is X + 2*Y),
        cont_f(N1, FNA, FNB, P),
        call(Pred, FNA, FNB, FN1, FN)
    ).

答案 3 :(得分:1)

记忆很有用,我把它用于使用Erathostenes筛子进行加速计算

?- time((between(0,1000,N), prob3(N,_), false)).
% 10,939 inferences, 0.011 CPU in 0.012 seconds (99% CPU, 951780 Lips)

:- dynamic memo/2.
prob3(0, 0).
prob3(1, 1).
prob3(N, R) :- memo(N, R), !.
prob3(N, R) :-
    N > 1, N2 is N-2, N1 is N-1, prob3(N2,R2), prob3(N1,R1), R is R2+2*R1, assertz(memo(N, R)).