There are dozens of ways of computing F(n) for an arbitrary n, many of which have great runtime and memory usage.
However, suppose I wanted to ask the opposite question:
Given F(n) for n > 2, what is n?
(The n > 2 restriction is in there since F(1) = F(2) = 1 and there's no unambiguous inverse).
What would be the most efficient way of solving this problem? It's easy to do this in linear time by enumerating the Fibonacci numbers and stopping when you hit the target number, but is there some way of doing this any faster than that?
EDIT: currently, the best solution posted here runs in O(log n) time using O(log n) memory, assuming that mathematical operations run in O(1) and that a machine word can hold any number in O(1) space. I'm curious if it's possible to drop the memory requirements, since you can compute Fibonacci numbers using O(1) space.
Since OP has asked about matrix solution not involving any floating point computations, here it is. We can achieve O(logn) complexity this way, assuming numeric operations have O(1) complexity.
Let's take 2x2 matrix A having following structure
1 1
1 0
Now consider vector (8, 5), storing two consecutive fibonacci numbers. If you multiply it by this matrix, you'll get (8*1 + 5*1, 8*1 + 5*0) = (13, 8) - the next fibonacci number.
If we generalize, A^n * (1, 0) = (f(n), f(n - 1)).
The actual algorithm takes two steps.
Calculate A^2, A^4, A^8, etc. until we pass desired number.
Do a binary search by n, using calculated powers of A.
On a side note, any sequence of the form f(n) = k1*f(n-1) + k2*f(n-2) + k3*f(n-3) + .. + kt*f(n-t) can be presented like this.
Wikipedia gives the result as
n(F) = Floor[ Log(F Sqrt(5) + 1/2)/Log(Phi)]
where Phi is the golden ratio.
If you can easily interpret F(n) in binary,
You may be suspicious of the constants 1.7 and 1.1. These work because d*1.44042009041... + C never gets very close to an integer.
I can post a derivation tomorrow if there is interest.
Here is a table with n = 2 through 91, which shows the formula result before flooring:
n formula w/o floor F(n) F(n) in binary
2 2.540 1 1
3 3.981 2 10
4 4.581 3 11
5 5.421 5 101
6 6.862 8 1000
7 7.462 13 1101
8 8.302 21 10101
9 9.743 34 100010
10 10.343 55 110111
11 11.183 89 1011001
12 12.623 144 10010000
13 13.223 233 11101001
14 14.064 377 101111001
15 15.504 610 1001100010
16 16.104 987 1111011011
17 17.545 1597 11000111101
18 18.385 2584 101000011000
19 19.825 4181 1000001010101
20 20.425 6765 1101001101101
21 21.266 10946 10101011000010
22 22.706 17711 100010100101111
23 23.306 28657 110111111110001
24 24.147 46368 1011010100100000
25 25.587 75025 10010010100010001
26 26.187 121393 11101101000110001
27 27.028 196418 101111111101000010
28 28.468 317811 1001101100101110011
29 29.068 514229 1111101100010110101
30 30.508 832040 11001011001000101000
31 31.349 1346269 101001000101011011101
32 32.789 2178309 1000010011110100000101
33 33.389 3524578 1101011100011111100010
34 34.230 5702887 10101110000010011100111
35 35.670 9227465 100011001100110011001001
36 36.270 14930352 111000111101000110110000
37 37.111 24157817 1011100001001111001111001
38 38.551 39088169 10010101000111000000101001
39 39.151 63245986 11110001010000111010100010
40 40.591 102334155 110000110010111111011001011
41 41.432 165580141 1001110111101000110101101101
42 42.032 267914296 1111111110000000110000111000
43 43.472 433494437 11001110101101001100110100101
44 44.313 701408733 101001110011101010010111011101
45 45.753 1134903170 1000011101001010011111110000010
46 46.353 1836311903 1101101011100111110010101011111
47 47.193 2971215073 10110001000110010010010011100001
48 48.634 4807526976 100011110100011010000101001000000
49 49.234 7778742049 111001111101001100010111100100001
50 50.074 12586269025 1011101110001100110011100101100001
51 51.515 20365011074 10010111101110110010110100010000010
52 52.115 32951280099 11110101100000011001010000111100011
53 53.555 53316291173 110001101001111001100000101001100101
54 54.396 86267571272 1010000010101111100101010110001001000
55 55.836 139583862445 10000001111111110110001011011010101101
56 56.436 225851433717 11010010010101110010110110001011110101
57 57.276 365435296162 101010100010101101001000001100110100010
58 58.717 591286729879 1000100110101011011011110111110010010111
59 59.317 956722026041 1101111011000001000100111001011000111001
60 60.157 1548008755920 10110100001101100100000110001001011010000
61 61.598 2504730781961 100100011100101101100101101010100100001001
62 62.198 4052739537881 111010111110011010000110011011101111011001
63 63.038 6557470319842 1011111011011000111101100000110010011100010
64 64.478 10610209857723 10011010011001100001110010100010000010111011
65 65.078 17167680177565 11111001110100101001011110101000010110011101
66 66.519 27777890035288 110010100001110001011010001001010011001011000
67 67.359 44945570212853 1010001110000010110100101111110010101111110101
68 68.800 72723460248141 10000100010010001000000000000111101001001001101
69 69.400 117669030460994 11010110000010011110100110000101111111001000010
70 70.240 190392490709135 101011010010100100110100110001101101000010001111
71 71.681 308061521170129 1000110000010111000101001100010011100111011010001
72 72.281 498454011879264 1110001010101011101011110010100001001111101100000
73 73.121 806515533049393 10110111011000010110000111110110100110111000110001
74 74.561 1304969544928657 100101000101101110011100110001010110000110110010001
75 75.161 2111485077978050 111100000000110001001101110000001010111101111000010
76 76.602 3416454622906707 1100001000110011111101010100001100001000100101010011
77 77.442 5527939700884757 10011101000111010000111000010001101100000010100010101
78 78.042 8944394323791464 11111110001101110000100010110011001101000111001101000
79 79.483 14472334024676221 110011011010101000001011011000100111001001001101111101
80 80.323 23416728348467685 1010011001100010110001111101111000000110010000111100101
81 81.764 37889062373143906 10000110100110111110011011000111100111111011010101100010
82 82.364 61305790721611591 11011001110011010100101010110110101000101101011101000111
83 83.204 99194853094755497 101100000011010010011000101111110010000101000110010101001
84 84.644 160500643816367088 1000111010001101100111110000110100111001010110001111110000
85 85.244 259695496911122585 1110011010100111111010110110110011001001111111000010011001
86 86.085 420196140727489673 10111010100110101100010100111101000000011010101010010001001
87 87.525 679891637638612258 100101101111011101011101011110011011001101010100010100100010
88 88.125 1100087778366101931 111101000100010011000000000110000011010000101001100110101011
89 89.566 1779979416004714189 1100010110011110000011101100100011110011101111101111011001101
90 90.406 2880067194370816120 10011111111000000011011101101010100001101110100111100001111000
91 91.846 4660046610375530309 100000010101011110011111011001111000000001100100101011101000101
Derivation
Required Precision for α = log 2/log Φ ≈ 1.44042...
Measuring memory usage by counting unbounded words is sort of silly, but as long as that's the model, there's an O(log n) time, O(1) word solution similar to Nikita Rybak's that in essence computes n via its Zeckendorf representation, which is based on the Fibonacci numbers (YO DAWG).
Define the matrix
1 1
A = ,
1 0
which satisfies
F(m + 1) F(m)
A^m = .
F(m) F(m - 1)
Instead of the sequence A^(2^k), we're going to use the sequence A^F(k). The latter sequence has the property that we can move forward with a matrix multiply
A^F(k + 1) = A^F(k - 1) * A^F(k)
and backward with a matrix inverse and multiplication
A^F(k - 1) = A^F(k + 1) (A^F(k))^-1,
so we can build a bidirectional iterator with only eight six twelve words assuming we store everything as rationals (to avoid assuming the existence of a unit-cost divide). The rest is just adapting this O(1)-space algorithm for finding a Zeckendorf representation.
def zeck(n):
a, b = (0, 1)
while b < n:
a, b = (b, a + b)
yield a
n1 = a
while n1 < n:
a, b = (b - a, a)
if n1 + a <= n:
yield a
n1 += a
a, b = (b - a, a)
>>> list(zeck(0))
[0]
>>> list(zeck(2))
[1, 1]
>>> list(zeck(12))
[8, 3, 1]
>>> list(zeck(750))
[610, 89, 34, 13, 3, 1]
It's been proven that the formula for a fib n is fib(n) = ( (phi)^n - (-phi)^(-n) ) / sqrt(5) where phi = (1+sqrt(5)) / 2, the golden section number. (see this link).
You could try to find a mathematical inverse to the fib function above, or otherwise do a binary search in 32/64 operations (depending on how big your searchable maximum is) to find the n that matches the number (try each n by computing fib(n) and splitting your sample space in two according to how fib(n) compares to the given fibonacci number).
Edit: #rcollyer's solution is faster, as mine is in O(lg n) and the one he found is in O(1) = constant time.
So I was thinking about this problem and I think that it's possible to do this in O(lg n) time with O(lg n) memory usage. This is based on the fact that
F(n) = (1 / √5) (Φn - φn)
Where Φ = (1 + √5)/2 and φ = 1 - Φ.
The first observation is that φn < 1 for any n > 1. This means that for any n > 2, we have that
F(n) = ⌊ Φn / √5 ⌋
Now, take n and write it in binary as bk-1bk-2...b1b0. This means that
n = 2k-1 bk-1 + 2k-2 bk-2 + ... + 21 b1 + 20 b0.
This means that
F(n) = ⌊ Φ2k-1 bk-1 + 2k-2 bk-2 + ... + 21 b1 + 20 b0 / √5 ⌋
Or, more readably, that
F(n) = ⌊ Φ2k-1 bk-1Φ2k-2 bk-2 ... Φ21 b1Φ20 b0 / √5 ⌋
This suggests the following algorithm. First, start computing Φ2k for all k until you compute a number Φz such that ⌊ Φz / √5 ⌋ that's greater than your number F(n). Now, from there, iterate backwards across all of the powers of Φ you generated this way. If the current number is bigger than the indicated power of Φ, then divide it by that power of Φ and record that the number was divided by this value. This process essentially recovers one bit of n at a time by subtracting out the largest power of 2 that you can at a time. Consequently, once you're done, you'll have found n.
The runtime of this algorithm is O(lg n), since you can generate Φ2i by repeated squaring, and we only generate O(lg n) terms. The memory usage is O(lg n), since we store all of these values.
You can find n for any Fib(n) in O(1) time and O(1) space.
You can use a fixed-point CORDIC algorithm to compute ln() using only shift and add on integer data types.
If x = Fib(n), then n can be determined by
n = int(2.0801 * ln(x) + 2.1408)
CORDIC run-time is determined by the desired level of precision. The two floating-point values would be encoded as fixed-point values.
The only issue with this proposal is that it returns a value for numbers that are not in the Fibonacci sequence, but the original problem specifically stated that the input to the function would be Fib(n), which implies that only valid Fibonacci numbers would be used.
EDIT: Never mind. The asker has stated in comments that exponentiation is definitely not constant time.
Is exponentiation one of the mathematical operations that you'll allow in constant time? If so, we can compute F(n) in constant time via the closed-form formula. Then, given some F, we can do the following:
Compute F(1), F(2), F(4), F(16), F(256), ... until F(2^k) <= F < F(2^{k+1})
Do a binary search for i between 2^k and 2^{k+1} until F(i) <= F < F(i+1)
If F = F(n), then first part takes k = O(log(n)) steps. The second part is a binary search over a range of size O(2^k), so it also takes k = O(log(n)). So, in total, we have O(log(n)) time in O(1) space if (and it's a big if) we have exponentiation in O(1) time.
A closed form of the Fibonacci number formula is:
Fn = Round(φ^n / Sqrt(5))
Where φ is the golden ratio.
If we ignore the rounding factor this is invertible and the inverse function is:
F(-1)n= log(n*Sqrt(5))/logφ
Because we ignored the rounding factor there is an error in the formula which could be calculated. However if we consider that a number n is a Fibonacci number iff the interval [n*φ - 1/n, n*φ + 1/n] contains a natural number then:
A number is a Fibonacci number iff the interval [n*φ - 1/n, n*φ + 1/n] contains a natural number and that number's index in the Fibonacci sequence is given by rounding log(n*Sqrt(5))/logφ
This should be doable in (pseudo)-constant time depending on the algorithms used for calculating the log and square roots etc.
Edit: φ = (1+Sqrt(5))/2
This might be similar to user635541's answer. I don't fully understand his approach.
Using the matrix representation for Fibonacci numbers, discussed in other answers, we get a way to go from F_n and F_m to F_{n+m} and F_{n-m} in constant time, using only plus, multiplication, minus and division (actually not! see the update). We also have a zero (the identity matrix), so it is a mathematical group!
Normally when doing binary search we also want a division operator for taking averages. Or at least division by 2. However if we want to go from F_{2n} to F_n it requires a square root. Luckily it turns out that plus and minus are all we need for a logarithmic time 'nearly' binary search!
Wikipedia writes about the approach, ironically called Fibonacci_search, but the article is not very clearly written, so I don't know if it is exactly the same approach as mine. It is very important to understand that the Fibonacci numbers used for the Fibonacci search have nothing to do with the numbers we are looking for. It's a bit confusing. To demonstrate the approach, here is first an implementation of standard 'binary search' only using plus and minus:
def search0(test):
# Standard binary search invariants:
# i <= lo then test(i)
# i >= hi then not test(i)
# Extra invariants:
# hi - lo = b
# a, b = F_{k-1}, F_k
a, b = 0, 1
lo, hi = 0, 1
while test(hi):
a, b = b, a + b
hi = b
while b != 1:
mi = lo + a
if test(mi):
lo = mi
a, b = 2*a - b, b - a
else:
hi = mi
a, b = b - a, a
return lo
>>> search0(lambda n: n**2 <= 25)
5
>>> search0(lambda n: 2**n <= 256)
8
Here test is some boolean function; a and b are consecutive fibonacci numbers f_k and f_{k-1} such that the difference between out upper bound hi and lower bound lo is always f_k. We need both a and b so we can increase and decrease the implicit variable k efficiently.
Alright, so how do we use this to solve the problem? I found it useful to create a wrapper around our Fibonacci representation, that hides the matrix details. In practice (is there such a thing for a Fibonacci searcher?) you would want to inline everything manually. That would spare you the redundancy in the matrices and make some optimization around the matrix inversion.
import numpy as np
class Fib:
def __init__(self, k, M):
""" `k` is the 'name' of the fib, e.g. k=6 for F_6=8.
We need this to report our result in the very end.
`M` is the matrix representation, that is
[[F_{k+1}, F_k], [F_k, F_{k-1}]] """
self.k = k
self.M = M
def __add__(self, other):
return Fib(self.k + other.k, self.M.dot(other.M))
def __sub__(self, other):
return self + (-other)
def __neg__(self):
return Fib(-self.k, np.round(np.linalg.inv(self.M)).astype(int))
def __eq__(self, other):
return self.k == other.k
def value(self):
return self.M[0,1]
However the code does work, so we can test it as follows. Notice how little different the search function is from when our objects were integers and not Fibonaccis.
def search(test):
Z = Fib(0, np.array([[1,0],[0,1]])) # Our 0 element
A = Fib(1, np.array([[1,1],[1,0]])) # Our 1 element
a, b = Z, A
lo, hi = Z, A
while test(hi.value()):
a, b = b, a + b
hi = b
while b != A:
mi = lo + a
if test(mi.value()):
lo = mi
a, b = a+a-b, b-a
else:
hi = mi
a, b = b-a, a
return lo.k
>>> search(lambda n: n <= 144)
12
>>> search(lambda n: n <= 0)
0
The remaining open question is whether there is an efficient search algorithm for monoids. That is one that doesn't need a minus / additive inverse. My guess is no: that without minus you need the extra memory of Nikita Rybak.
Update
I just realized that we don't need division at all. The determinant of the F_n matrix is just (-1)^n, so we can actually do everything without division. In the below I removed all the matrix code, but I kept the Fib class, just because everything got so extremely messy otherwise.
class Fib2:
def __init__(self, k, fp, f):
""" `fp` and `f` are F_{k-1} and F_{k} """
self.k, self.fp, self.f = k, fp, f
def __add__(self, other):
fnp, fn, fmp, fm = self.fp, self.f, other.fp, other.f
return Fib2(self.k + other.k, fn*fm+fnp*fmp, (fn+fnp)*fm+fn*fmp)
def __sub__(self, other):
return self + (-other)
def __neg__(self):
fp, f = self.f + self.fp, -self.f
return Fib2(-self.k, (-1)**self.k*fp, (-1)**self.k*f)
def __eq__(self, other):
return self.k == other.k
def value(self):
return self.f
def search2(test):
Z = Fib2(0, 1, 0)
A = Fib2(1, 0, 1)
...
>>> search2(lambda n: n <= 280571172992510140037611932413038677189525)
200
>>> search2(lambda n: n <= 4224696333392304878706725602341482782579852840250681098010280137314308584370130707224123599639141511088446087538909603607640194711643596029271983312598737326253555802606991585915229492453904998722256795316982874482472992263901833716778060607011615497886719879858311468870876264597369086722884023654422295243347964480139515349562972087652656069529806499841977448720155612802665404554171717881930324025204312082516817125)
2000
>>> search2(lambda n: n <= 2531162323732361242240155003520607291766356485802485278951929841991312781760541315230153423463758831637443488219211037689033673531462742885329724071555187618026931630449193158922771331642302030331971098689235780843478258502779200293635651897483309686042860996364443514558772156043691404155819572984971754278513112487985892718229593329483578531419148805380281624260900362993556916638613939977074685016188258584312329139526393558096840812970422952418558991855772306882442574855589237165219912238201311184749075137322987656049866305366913734924425822681338966507463855180236283582409861199212323835947891143765414913345008456022009455704210891637791911265475167769704477334859109822590053774932978465651023851447920601310106288957894301592502061560528131203072778677491443420921822590709910448617329156135355464620891788459566081572824889514296350670950824208245170667601726417091127999999941149913010424532046881958285409468463211897582215075436515584016297874572183907949257286261608612401379639484713101138120404671732190451327881433201025184027541696124114463488665359385870910331476156665889459832092710304159637019707297988417848767011085425271875588008671422491434005115288334343837778792282383576736341414410248994081564830202363820504190074504566612515965134665683289356188727549463732830075811851574961558669278847363279870595320099844676879457196432535973357128305390290471349480258751812890314779723508104229525161740643984423978659638233074463100366500571977234508464710078102581304823235436518145074482824812996511614161933313389889630935320139507075992100561077534028207257574257706278201308302642634678112591091843082665721697117838726431766741158743554298864560993255547608496686850185804659790217122426535133253371422250684486113457341827911625517128815447325958547912113242367201990672230681308819195941016156001961954700241576553750737681552256845421159386858399433450045903975167084252876848848085910156941603293424067793097271128806817514906531652407763118308162377033463203514657531210413149191213595455280387631030665594589183601575340027172997222489081631144728873621805528648768511368948639522975539046995395707688938978847084621586473529546678958226255042389998718141303055036060772003887773038422366913820397748550793178167220193346017430024134496141145991896227741842515718997898627269918236920453493946658273870473264523119133765447653295022886429174942653014656521909469613184983671431465934965489425515981067546087342348350724207583544436107294087637975025147846254526938442435644928231027868701394819091132912397475713787593612758364812687556725146456646878912169274219209708166678668152184941578590201953144030519381922273252666652671717526318606676754556170379350956342095455612780202199922615392785572481747913435560866995432578680971243966868110016581395696310922519803685837460795358384618017215468122880442252343684547233668502313239328352671318130604247460452134121833305284398726438573787798499612760939462427922917659263046333084007208056631996856315539698234022953452211505675629153637867252695056925345220084020071611220575700841268302638995272842160994219632684575364180160991884885091858259996299627148614456696661412745040519981575543804847463997422326563897043803732970397488471644906183310144691243649149542394691524972023935190633672827306116525712882959108434211652465621144702015336657459532134026915214509960877430595844287585350290234547564574848753110281101545931547225811763441710217452979668178025286460158324658852904105792472468108996135476637212057508192176910900422826969523438985332067597093454021924077101784215936539638808624420121459718286059401823614213214326004270471752802725625810953787713898846144256909835116371235019527013180204030167601567064268573820697948868982630904164685161783088076506964317303709708574052747204405282785965604677674192569851918643651835755242670293612851920696732320545562286110332140065912751551110134916256237884844001366366654055079721985816714803952429301558096968202261698837096090377863017797020488044826628817462866854321356787305635653577619877987998113667928954840972022833505708587561902023411398915823487627297968947621416912816367516125096563705174220460639857683971213093125)
20000
This all works like a charm. My only worry is that the bit complexity such dominates the calculation, that we might as well have just done a sequential search. Or actually, just looking at the number of digits could probably tell you pretty much which you were looking at. That's not as fun though.
Related
I'm solving some Project Euler problems using Ruby, and specifically here I'm talking about problem 25 (What is the index of the first term in the Fibonacci sequence to contain 1000 digits?).
At first, I was using Ruby 2.2.3 and I coded the problem as such:
number = 3
a = 1
b = 2
while b.to_s.length < 1000
a, b = b, a + b
number += 1
end
puts number
But then I found out that version 2.4.2 has a method called digits which is exactly what I needed. I transformed to code to:
while b.digits.length < 1000
And when I compared the two methods, digits was much slower.
Time
./025/problem025.rb 0.13s user 0.02s system 80% cpu 0.190 total
./025/problem025.rb 2.19s user 0.03s system 97% cpu 2.275 total
Does anyone have an idea why?
Ruby's digits
... is implemented in rb_int_digits.
Which for non-tiny numbers (i.e., most of your numbers) uses rb_int_digits_bigbase.
Which extracts digit after digit naively with division/modulo by base.
So it should take quadratic time (at least with a small base such as 10).
Ruby's to_s
... is implemented in int_to_s.
Which uses rb_int2str.
Which for non-tiny numbers uses rb_big2str.
Which uses rb_big2str1.
Which might use big2str_gmp if available (which sounds/looks like it uses the fast GMP library) or ...
... uses big2str_generic.
Which uses big2str_karatsuba (sweet, I recognize that name!).
Which looks like it has something to do with ...
... Karatsuba's algorithm, which is a fast multiplication algorithm. If you multiply two n-digit numbers the naive way you learned in school, you take n2 single-digit products. Karatsuba on the other hand only needs about n1.585, which is quite a lot better. And I didn't read into this further, but I suspect what Ruby does here is also this efficient. Eric Lippert's answer with a base conversion algorithm uses Karatsuba multiplication and says "this [base conversion] algorithm is utterly dominated by the cost of the multiplication".
Comparing quadratic to n1.585 over the number lengths from 1 digit to 1000 digits gives factor 15:
(1..1000).sum { |i| i**2 } / (1..1000).sum { |i| i**1.585 }
=> 15.150583254950678
Which is roughly the factor you observed as well. Of course that's a rather naive comparison, but, well, why not.
GMP by the way apparently uses/used a "near O(n * log(n)) FFT-based multiplication algorithm".
Thanks to #Drenmi's answer for motivating me to dig into the source after all. I hope I did this right, no guarantees, I'm a Ruby beginner. But that's why I left all the links there for you to check for yourself :-P
Integer#digits doesn't just "split" the number. From the documentation:
Returns the array including the digits extracted by place-value
notation with radix base of int.
This extraction is done even if a base argument is omitted. The relevant source:
# ruby/numeric.c:4809
while (!FIXNUM_P(num) || FIX2LONG(num) > 0) {
VALUE qr = rb_int_divmod(num, base);
rb_ary_push(digits, RARRAY_AREF(qr, 1));
num = RARRAY_AREF(qr, 0);
}
As you can see, this process includes repeated modulo arithmetics, which likely accounts for the additional runtime.
Many ruby methods create objects (strins, arrays, etc.)
In ruby, object creation in ruby is "expensive".
For instance to_s creates a string and digits creates an array every time the while condition is evaluated.
If you want to optimize your example, you can do the following:
# create the smallest possible 1000 digits number
max = 10**999
number = 3
a = 1
b = 2
# do not create objects in while condition
while b < max
a, b = b, a + b
number += 1
end
puts number
I have not answered your question, but wish to suggest an improved algorithm for the problem you have addressed. For a given number of decimal digits, n, I have implemented the following algorithm.
estimate the number f of Fibonacci numbers ("FNs") that have n or fewer decimal digits.
compute the fth and (f-1)st FNs, and the number of digits m in the fth FN.
if m >= n back down from down from the (f-1)st FN until the (f-1)st FN has fewer than n decimal digits, at which time the fth FN is the smallest FN to have n decimal digits.
if m < n increase the fth FN until the it has n decimal digits, at which time it is the smallest FN to have n decimal digits.
The key is to compute a close estimate f in the first step.
Code
AVG_FNs_PER_DIGIT = 4.784971966781667
def first_fibonacci_with_n_digits(n)
return [1, 1] if n == 1
idx = (n * AVG_FNs_PER_DIGIT).round
fn, prev_fn = fib(idx)
fn.to_s.size >= n ? fib_down(n, fn, prev_fn, idx) : fib_up(n, fn, prev_fn, idx)
end
def fib(idx)
a = 1
b = 2
(idx - 2).times {a, b = b, a + b }
[b, a]
end
def fib_up(n, b, a, idx)
loop do
a, b = b, a + b
idx += 1
break [idx, b] if b.to_s.size == n
end
end
def fib_down(n, b, a, idx)
loop do
a, b = b - a, a
break [idx, b] if a.to_s.size == n - 1
idx -= 1
end
end
Benchmarks
In computing each Fibonacci number two operations are typically performed:
compute the number of digits in the last-computed Fibonacci number and if that number is equal to the target number of digits, terminate (for reasons made clear in the Explanation section below, it cannot be larger than the target number); else
compute the next number in the Fibonacci sequence.
By contrast, the method I have proposed performs the first step a relatively small number of times.
How important is the first step relative to the second and how does the use of n.digits.size compare with that of n.to_s.size in the first step? Let's run some benchmarks to find out.
def use_to_s(ndigits)
case ndigits
when 1
[1, 1]
else
a = 1
b = 2
idx = 3
loop do
break [idx, b] if b.to_s.length == ndigits
a, b = b, a + b
idx += 1
end
end
end
def use_digits(ndigits)
case ndigits
when 1
[1, 1]
else
a = 1
b = 2
idx = 3
loop do
break [idx, b] if b.digits.size == ndigits
a, b = b, a + b
idx += 1
end
end
end
require 'fruity'
def test(ndigits)
nfibs, last_fib = use_to_s(ndigits)
puts "\nndigits = #{ndigits}, nfibs=#{nfibs}, last_fib=#{last_fib}"
compare do
try_use_to_s { use_to_s(ndigits) }
try_use_digits { use_digits(ndigits) }
try_estimate { first_fibonacci_with_n_digits(ndigits) }
end
end
test 20
ndigits = 20, nfibs=93, last_fib=12200160415121876738
Running each test 128 times. Test will take about 1 second.
try_estimate is faster than try_use_to_s by 2x ± 0.1
try_use_to_s is faster than try_use_digits by 80.0% ± 10.0%
test 100
ndigits = 100, nfibs=476, last_fib=13447...37757 (90 digits omitted)
Running each test 16 times. Test will take about 4 seconds.
try_estimate is faster than try_use_to_s by 5x ± 0.1
try_use_to_s is faster than try_use_digits by 10x ± 1.0
test 500
ndigits = 500, nfibs=2390, last_fib=13519...63145 (490 digits omitted)
Running each test 2 times. Test will take about 27 seconds.
try_estimate is faster than try_use_to_s by 9x ± 0.1
try_use_to_s is faster than try_use_digits by 60x ± 1.0
test 1000
ndigits = 1000, nfibs=4782, last_fib=10700...27816 (990 digits omitted)
Running each test once. Test will take about 1 minute.
try_estimate is faster than try_use_to_s by 12x ± 10.0
try_use_to_s is faster than try_use_digits by 120x ± 100.0
There are two main take-aways from these results:
"try_estimate" is the fastest because it performs the first step relatively few times; and
the use of to_s is much faster than that of digits.
Further to the first of these observations note that the initial estimates of the index of the first FN having a given number of digits, compared to the actual index, are as follows:
for 20 digits: 96 est. vs 93 actual
for 100 digits: 479 est. vs 476 actual
for 500 digits: 2392 est. vs 2390 actual
for 1000 digits: 4785 est. vs 4782 actual
The deviation was at most 3, meaning numbers of digits had to be calculated for at most 3 FNs to obtain the desired result.
Explanation
The only explanation of the methods given in the section Code above is the derivation of the constant AVG_FNs_PER_DIGIT, which is used to calculate an estimate of the index of the first FN having the specified number of digits.
The derivation of this constant derives from the question and selected answer given here. (The Wiki for Fibonacci numbers provides a good overview of the mathematical properties of FNs.)
It is known that the first 7 FNs (including zero) have one digit; thereafter the FNs gain an additional digit every 4 or 5 FNs (i.e., sometimes 4, else 5). Therefore, as a very crude calculation, we see that to calculate the first FN with n digits, n >= 2, it will not be less than the 4*nth FN. For n = 1000, that would be 4,000. (In fact, the 4,782nd is the smallest to have 1,000 digits.) In other words, we don't need to calculate the number of digits in the first 4,000 FNs. We can improve on this estimate, however.
As n approaches infinity, the ratio of ranges 10**n...10**(n+1) (n-digit intervals) that contain 5 FNs to those that contain 4 FNs can be computed as follows.
LOG_10 = Math.log(10)
#=> 2.302585092994046
GR = (1 + Math.sqrt(5))/2
#=> 1.618033988749895
LOG_GR = Math.log(GR)
#=> 0.48121182505960347
RATIO_5to4 = (LOG_10 - 4*LOG_GR)/(5*LOG_GR - LOG_10)
#=> 3.6505564183095474
where GR is the Golden Ratio.
Over a large number of n-digit intervals let n4 be the number of those intervals containing 4 FNs and n5 be the number containing 5 FNs. The average number of FNs per interval is therefore (n4*4 + n5*5)/(n4 + n5). Since n5/n4 converges to RATIO_5to4, n5 approaches RATIO_5to4 * n4 in the limit (discarding roundoff error). If we substitute out n5, and let
b = 1/(1 + RATIO_5to4)
#=> 0.21502803321833364
we find the average number of FNs per n-digit interval converges to
avg = b * 4 + (1-b) *5
#=> 4.784971966781667
If fn is the first FN to have n decimal digits, the number of FNs in the sequence up to an including fn can therefore be approximated to be
n * avg
If, for example, the estimate of the index of the first FN to have 1000 decimal digits would be 1000 * 4.784971966781667).round #=> 4785.
I am solving this problem where we need to reach from X=0 to X=N.We can only take a step of 2 or 3 at a time.
For each step of 2 we have a probability of 0.2 and for each step of 3 we have a probability of 0.8.How can we find the total probability to reach N.
e.g. for reaching 5,
2+3 with probability =0.2 * 0.8=0.16
3+2 with probability =0.8 * 0.2=0.16 total = 0.32.
My initial thoughts:
Number of ways can be found out by simple Fibonacci sequence.
f(n)=f(n-3)+f(n-2);
But how do we remember the numbers so that we can multiply them to find the probability?
This can be solved using Dynamic programming.
Lets call the function F(N) = probability to reach 0 using only 2 and 3 when the starting number is N
F(N) = 0.2*F(N-2) + 0.3*F(N-3)
Base case:
F(0) = 1 and F(k)= 0 where k< 0
So the DP code would be somthing like that:
F[0] = 1;
for(int i = 1;i<=N;i++){
if(i>=3)
F[i] = 0.2*F[i-2] + 0.8*F[i-3];
else if(i>=2)
F[i] = 0.2*F[i-2];
else
F[i] = 0;
}
return F[N];
This algorithm would run in O(N)
Some clarifications about this solution: I assume the only allowed operation for generating the number from 2s and 3s is addition (your definition would allow substraction aswell) and the input-numbers are always valid (2 <= input). Definition: a unique row of numbers means: no other row with the same number of 3s and 2s in another order is in scope.
We can reduce the problem into multiple smaller problems:
Problem A: finding all sequences of numbers that can sum up to the given number. (Unique rows of numbers only)
Start by finding the minimum-number of 3s required to build the given number, which is simply input % 2. The maximum-number of 3s that can be used to build the input can be calculated this way:
int max_3 = (int) (input / 3);
if(input - max_3 == 1)
--max_3;
Now all sequences of numbers that sum up to input must hold between input % 2 and max_3 3s. The 2s can be easily calculated from a given number of 3s.
Problem B: calculating the probability for a given list and it's permutations to be the result
For each unique row of numbers, we can easily derive all permutations. Since these consist of the same number, they have the same likeliness to appear and produce the same sum. The likeliness can be calculated easily from the row: 0.8 ^ number_of_3s * 0.2 ^ number_of_2s. Next step would be to calculate the number of different permuatations. Calculating all distinct sets with a specific number of 2s and 3s can be done this way: Calculate all possible distributions of 2s in the set: (number_of_2s + number_of_3s)! / (number_of_3s! * numer_of_2s!). Basically just the number of possible distinct permutations.
Now from theory to praxis
Since the math is given, the rest is pretty straight forward:
define prob:
input: int num
output: double
double result = 0.0
int min_3s = (num % 2)
int max_3s = (int) (num / 3)
if(num - max_3 == 1)
--max_3
for int c3s in [min_3s , max_3s]
int c2s = (num - (c3s * 3)) / 2
double p = 0.8 ^ c3s * 0.2 * c2s
p *= (c3s + c2s)! / (c3s! * c2s!)
result += p
return result
Instead of jumping into the programming, you can use math.
Let p(n) be the probability that you reach the location that is n steps away.
Base cases:
p(0)=1
p(1)=0
p(2)=0.2
Linear recurrence relation
p(n+3)=0.2 p(n+1) + 0.8 p(n)
You can solve this in closed form by finding the exponential solutions to the linear recurrent relation.
c^3 = 0.2 c + 0.8
c = 1, (-5 +- sqrt(55)i)/10
Although this was cubic, c=1 will always be a solution in this type of problem since there is a constant nonzero solution.
Because the roots are distinct, all solutions are of the form a1(1)^n + a2((-5+sqrt(55)i) / 10)^n + a3((-5-sqrt(55)i)/10)^n. You can solve for a1, a2, and a3 using the initial conditions:
a1=5/14
a2=(99-sqrt(55)i)/308
a3=(99+sqrt(55)i)/308
This gives you a nonrecursive formula for p(n):
p(n)=5/14+(99-sqrt(55)i)/308((-5+sqrt(55)i)/10)^n+(99+sqrt(55)i)/308((-5-sqrt(55)i)/10)^n
One nice property of the non-recursive formula is that you can read off the asymptotic value of 5/14, but that's also clear because the average value of a jump is 2(1/5)+ 3(4/5) = 14/5, and you almost surely hit a set with density 1/(14/5) of the integers. You can use the magnitudes of the other roots, 2/sqrt(5)~0.894, to see how rapidly the probabilities approach the asymptotics.
5/14 - (|a2|+|a3|) 0.894^n < p(n) < 5/14 + (|a2|+|a3|) 0.894^n
|5/14 - p(n)| < (|a2|+|a3|) 0.894^n
f(n, p) = f(n-3, p*.8) + f(n -2, p*.2)
Start p at 1.
If n=0 return p, if n <0 return 0.
Instead of using the (terribly inefficient) recursive algorithm, start from the start and calculate in how many ways you can reach subsequent steps, i.e. using 'dynamic programming'. This way, you can easily calculate the probabilities and also have a complexity of only O(n) to calculate everything up to step n.
For each step, memorize the possible ways of reaching that step, if any (no matter how), and the probability of reaching that step. For the zeroth step (the start) this is (1, 1.0).
steps = [(1, 1.0)]
Now, for each consecutive step n, get the previously computed possible ways poss and probability prob to reach steps n-2 and n-3 (or (0, 0.0) in case of n < 2 or n < 3 respectively), add those to the combined possibilities and probability to reach that new step, and add them to the list.
for n in range(1, 10):
poss2, prob2 = steps[n-2] if n >= 2 else (0, 0.0)
poss3, prob3 = steps[n-3] if n >= 3 else (0, 0.0)
steps.append( (poss2 + poss3, prob2 * 0.2 + prob3 * 0.8) )
Now you can just get the numbers from that list:
>>> for n, (poss, prob) in enumerate(steps):
... print "%s\t%s\t%s" % (n, poss, prob)
0 1 1.0
1 0 0.0
2 1 0.2
3 1 0.8
4 1 0.04
5 2 0.32 <-- 2 ways to get to 5 with combined prob. of 0.32
6 2 0.648
7 3 0.096
8 4 0.3856
9 5 0.5376
(Code is in Python)
Note that this will get you both the number of possible ways of reaching a certain step (e.g. "first 2, then 3" or "first 3, then 2" for 5), and the probability to reach that step in one go. Of course, if you need only the probability, you can just use single numbers instead of tuples.
So I have a brain teaser I read on one of the algorithm and puzzle meetups we have on our uni that goes like this:
There's a school that awards students that, during a given period, are
never late more than once and who don't ever happen to be absent for
three or more consecutive days. How many possible permutations with repetitions of
presence (or lack thereof) can we build for a given timeframe that
grant the student an award? Assume that each day is just a state
On-time, Late or Absent for the whole day, don't worry about specific
classes. Example: for three day timeframes, we can create 19 such
permutations with repetitions that grant an award.
I've already posted it on math.SE yesterday cause I was interested if there was some ready-bake formula we could derive to solve it but it turns out there isn't and all the transformations really are rather complex.
Thus, I'm asking here - how would you approach such a problem with an algorithm? I tried narrowing down the possibilities space but after a while taking all the possible permutations with repetitions became well too much and the algorithm started becoming really complex while I believe there should be some easy to implement way to solve it, especially since most of the puzzles we exchange on the meetup are rather like that.
Here is a simplified version of Python 3 code implementing the recursion in the answer by #ProgrammerPerson:
from functools import lru_cache
def count_variants(max_late, base_absent, period_length):
"""
max_late – maximum allowed number of days the student can be late;
base_absent – the number of consecutive days the student can be absent;
period_length – days in a period."""
#lru_cache(max_late * base_absent * period_length)
def count(late, absent, days):
if late < 0: return 0
if absent < 0: return 0
if days == 0: return 1
return (count(late, base_absent, days-1) + # Student is on time. Absent reset.
count(late-1, base_absent, days-1) + # Student is late. Absent reset.
count(late, absent-1, days-1)) # Student is absent.
return count(max_late, base_absent, period_length)
Run example:
In [2]: count_variants(1, 2, 3)
Out[2]: 19
This screams recursion (and/or dynamic programming)!
Suppose we try and solve a slightly general problem:
We give an award if a student is late no more than L times, and isn't
absent for A or more consecutive days.
Now we want to compute the number of possibilities for an n days time frame.
Call this method P(L, A, n)
Now try to build up a recursion based on three cases for the first day of the period.
1) If the student is on-time for the first day, then the number is simply
P(L, A, n-1)
2) If the student is late the first day, then the number is
P(L-1, A, n-1)
3) If the student is absent the first day, then the number is
P(L, A-1, n-1)
This gives us the recursion:
P(L, A, n) = P(L, A, n-1) + P(L-1, A, n-1) + P(L, A-1, n-1)
You can either memoize the recursion, or just have tables which you lookup.
Be careful about the base cases which are
P(0, *, *), P(*, 0, *) and P(*, *, 0) and can be computed by easy mathematical formulae.
Here is quick python code, with memoization + recursion to demonstrate:
import math
def binom(n, r):
return math.factorial(n)/(math.factorial(r)*math.factorial(n-r))
# The memoization table.
table = {}
def P(L, A, n):
if L == 0:
# Only ontime or absent.
# More absents than period.
if A > n:
return 2**n
# 2^n total possibilities.
# of that n-A+1 are non-rewarding.
return 2**n - (n - A + 1)
if A == 0:
# Only Late or ontime.
# need fewer than L+1 late.
# This is n choose 0 + n choose 1 + ... + n choose L
total = 0
for l in xrange(0, min(L,n)):
total += binom(n, l)
return total
if n == 0:
return 1
if (L, A, n) in table:
return table[(L, A, n)]
result = P(L, A, n-1) + P(L-1, A, n-1) + P(L, A-1, n-1)
table[(L, A, n)] = result
return result
print P(1, 3, 3)
Output is 19.
Let S(n) be the number of strings of length n without 3 repeated 1s.
Any such string (with length at least 3) ends in "0", "01" or "011" (and after removing the suffix, any string without three consecutive 1s can appear).
Then for n > 2, S(n) = S(n-1) + S(n-2) + S(n-3), and S(0)=1, S(1)=2, S(2)=4.
If you have a late day on day i (counting from 0), then you have S(i) ways of arranging absent days before, and S(n-i-1) ways of arranging absent days after.
Thus, the solution to the original problem is S(n) + sum(S(i)*S(n-i-1) | i = 0...n-1)
We can compute solutions iteratively like this:
def ways(n):
S = [1, 2, 4] + [0] * (n-2)
for i in xrange(3, n+1):
S[i] = S[i-1] + S[i-2] + S[i-3]
return S[n] + sum(S[i] * S[n-i-1] for i in xrange(n))
for i in xrange(1, 20):
print i, ways(i)
Output:
1 3
2 8
3 19
4 43
5 94
6 200
7 418
8 861
9 1753
10 3536
11 7077
12 14071
13 27820
14 54736
15 107236
16 209305
17 407167
18 789720
19 1527607
I recently encountered a much more difficult variation of this problem, but realized I couldn't generate a solution for this very simple case. I searched Stack Overflow but couldn't find a resource that previously answered this.
You are given a triangle ABC, and you must compute the number of paths of certain length that start at and end at 'A'. Say our function f(3) is called, it must return the number of paths of length 3 that start and end at A: 2 (ABA,ACA).
I'm having trouble formulating an elegant solution. Right now, I've written a solution that generates all possible paths, but for larger lengths, the program is just too slow. I know there must be a nice dynamic programming solution that reuses sequences that we've previously computed but I can't quite figure it out. All help greatly appreciated.
My dumb code:
def paths(n,sequence):
t = ['A','B','C']
if len(sequence) < n:
for node in set(t) - set(sequence[-1]):
paths(n,sequence+node)
else:
if sequence[0] == 'A' and sequence[-1] == 'A':
print sequence
Let PA(n) be the number of paths from A back to A in exactly n steps.
Let P!A(n) be the number of paths from B (or C) to A in exactly n steps.
Then:
PA(1) = 1
PA(n) = 2 * P!A(n - 1)
P!A(1) = 0
P!A(2) = 1
P!A(n) = P!A(n - 1) + PA(n - 1)
= P!A(n - 1) + 2 * P!A(n - 2) (for n > 2) (substituting for PA(n-1))
We can solve the difference equations for P!A analytically, as we do for Fibonacci, by noting that (-1)^n and 2^n are both solutions of the difference equation, and then finding coefficients a, b such that P!A(n) = a*2^n + b*(-1)^n.
We end up with the equation P!A(n) = 2^n/6 + (-1)^n/3, and PA(n) being 2^(n-1)/3 - 2(-1)^n/3.
This gives us code:
def PA(n):
return (pow(2, n-1) + 2*pow(-1, n-1)) / 3
for n in xrange(1, 30):
print n, PA(n)
Which gives output:
1 1
2 0
3 2
4 2
5 6
6 10
7 22
8 42
9 86
10 170
11 342
12 682
13 1366
14 2730
15 5462
16 10922
17 21846
18 43690
19 87382
20 174762
21 349526
22 699050
23 1398102
24 2796202
25 5592406
26 11184810
27 22369622
28 44739242
29 89478486
The trick is not to try to generate all possible sequences. The number of them increases exponentially so the memory required would be too great.
Instead, let f(n) be the number of sequences of length n beginning and ending A, and let g(n) be the number of sequences of length n beginning with A but ending with B. To get things started, clearly f(1) = 1 and g(1) = 0. For n > 1 we have f(n) = 2g(n - 1), because the penultimate letter will be B or C and there are equal numbers of each. We also have g(n) = f(n - 1) + g(n - 1) because if a sequence ends begins A and ends B the penultimate letter is either A or C.
These rules allows you to compute the numbers really quickly using memoization.
My method is like this:
Define DP(l, end) = # of paths end at end and having length l
Then DP(l,'A') = DP(l-1, 'B') + DP(l-1,'C'), similar for DP(l,'B') and DP(l,'C')
Then for base case i.e. l = 1 I check if the end is not 'A', then I return 0, otherwise return 1, so that all bigger states only counts those starts at 'A'
Answer is simply calling DP(n, 'A') where n is the length
Below is a sample code in C++, you can call it with 3 which gives you 2 as answer; call it with 5 which gives you 6 as answer:
ABCBA, ACBCA, ABABA, ACACA, ABACA, ACABA
#include <bits/stdc++.h>
using namespace std;
int dp[500][500], n;
int DP(int l, int end){
if(l<=0) return 0;
if(l==1){
if(end != 'A') return 0;
return 1;
}
if(dp[l][end] != -1) return dp[l][end];
if(end == 'A') return dp[l][end] = DP(l-1, 'B') + DP(l-1, 'C');
else if(end == 'B') return dp[l][end] = DP(l-1, 'A') + DP(l-1, 'C');
else return dp[l][end] = DP(l-1, 'A') + DP(l-1, 'B');
}
int main() {
memset(dp,-1,sizeof(dp));
scanf("%d", &n);
printf("%d\n", DP(n, 'A'));
return 0;
}
EDITED
To answer OP's comment below:
Firstly, DP(dynamic programming) is always about state.
Remember here our state is DP(l,end), represents the # of paths having length l and ends at end. So to implement states using programming, we usually use array, so DP[500][500] is nothing special but the space to store the states DP(l,end) for all possible l and end (That's why I said if you need a bigger length, change the size of array)
But then you may ask, I understand the first dimension which is for l, 500 means l can be as large as 500, but how about the second dimension? I only need 'A', 'B', 'C', why using 500 then?
Here is another trick (of C/C++), the char type indeed can be used as an int type by default, which value is equal to its ASCII number. And I do not remember the ASCII table of course, but I know that around 300 will be enough to represent all the ASCII characters, including A(65), B(66), C(67)
So I just declare any size large enough to represent 'A','B','C' in the second dimension (that means actually 100 is more than enough, but I just do not think that much and declare 500 as they are almost the same, in terms of order)
so you asked what DP[3][1] means, it means nothing as the I do not need / calculate the second dimension when it is 1. (Or one can think that the state dp(3,1) does not have any physical meaning in our problem)
In fact, I always using 65, 66, 67.
so DP[3][65] means the # of paths of length 3 and ends at char(65) = 'A'
You can do better than the dynamic programming/recursion solution others have posted, for the given triangle and more general graphs. Whenever you are trying to compute the number of walks in a (possibly directed) graph, you can express this in terms of the entries of powers of a transfer matrix. Let M be a matrix whose entry m[i][j] is the number of paths of length 1 from vertex i to vertex j. For a triangle, the transfer matrix is
0 1 1
1 0 1.
1 1 0
Then M^n is a matrix whose i,j entry is the number of paths of length n from vertex i to vertex j. If A corresponds to vertex 1, you want the 1,1 entry of M^n.
Dynamic programming and recursion for the counts of paths of length n in terms of the paths of length n-1 are equivalent to computing M^n with n multiplications, M * M * M * ... * M, which can be fast enough. However, if you want to compute M^100, instead of doing 100 multiplies, you can use repeated squaring: Compute M, M^2, M^4, M^8, M^16, M^32, M^64, and then M^64 * M^32 * M^4. For larger exponents, the number of multiplies is about c log_2(exponent).
Instead of using that a path of length n is made up of a path of length n-1 and then a step of length 1, this uses that a path of length n is made up of a path of length k and then a path of length n-k.
We can solve this with a for loop, although Anonymous described a closed form for it.
function f(n){
var as = 0, abcs = 1;
for (n=n-3; n>0; n--){
as = abcs - as;
abcs *= 2;
}
return 2*(abcs - as);
}
Here's why:
Look at one strand of the decision tree (the other one is symmetrical):
A
B C...
A C
B C A B
A C A B B C A C
B C A B B C A C A C A B B C A B
Num A's Num ABC's (starting with first B on the left)
0 1
1 (1-0) 2
1 (2-1) 4
3 (4-1) 8
5 (8-3) 16
11 (16-5) 32
Cleary, we can't use the strands that end with the A's...
You can write a recursive brute force solution and then memoize it (aka top down dynamic programming). Recursive solutions are more intuitive and easy to come up with. Here is my version:
# search space (we have triangle with nodes)
nodes = ["A", "B", "C"]
#cache # memoize!
def recurse(length, steps):
# if length of the path is n and the last node is "A", then it's
# a valid path and we can count it.
if length == n and ((steps-1)%3 == 0 or (steps+1)%3 == 0):
return 1
# we don't want paths having len > n.
if length > n:
return 0
# from each position, we have two possibilities, either go to next
# node or previous node. Total paths will be sum of both the
# possibilities. We do this recursively.
return recurse(length+1, steps+1) + recurse(length+1, steps-1)
Given: set A = {a0, a1, ..., aN-1} (1 ≤ N ≤ 100), with 2 ≤ ai ≤ 500.
Asked: Find the sum of all least common multiples (LCM) of all subsets of A of size at least 2.
The LCM of a setB = {b0, b1, ..., bk-1} is defined as the minimum integer Bmin such that bi | Bmin, for all 0 ≤ i < k.
Example:
Let N = 3 and A = {2, 6, 7}, then:
LCM({2, 6}) = 6
LCM({2, 7}) = 14
LCM({6, 7}) = 42
LCM({2, 6, 7}) = 42
----------------------- +
answer 104
The naive approach would be to simply calculate the LCM for all O(2N) subsets, which is not feasible for reasonably large N.
Solution sketch:
The problem is obtained from a competition*, which also provided a solution sketch. This is where my problem comes in: I do not understand the hinted approach.
The solution reads (modulo some small fixed grammar issues):
The solution is a bit tricky. If we observe carefully we see that the integers are between 2 and 500. So, if we prime factorize the numbers, we get the following maximum powers:
2 8
3 5
5 3
7 3
11 2
13 2
17 2
19 2
Other than this, all primes have power 1. So, we can easily calculate all possible states, using these integers, leaving 9 * 6 * 4 * 4 * 3 * 3 * 3 * 3 states, which is nearly 70000. For other integers we can make a dp like the following: dp[70000][i], where i can be 0 to 100. However, as dp[i] is dependent on dp[i-1], so dp[70000][2] is enough. This leaves the complexity to n * 70000 which is feasible.
I have the following concrete questions:
What is meant by these states?
Does dp stand for dynamic programming and if so, what recurrence relation is being solved?
How is dp[i] computed from dp[i-1]?
Why do the big primes not contribute to the number of states? Each of them occurs either 0 or 1 times. Should the number of states not be multiplied by 2 for each of these primes (leading to a non-feasible state space again)?
*The original problem description can be found from this source (problem F). This question is a simplified version of that description.
Discussion
After reading the actual contest description (page 10 or 11) and the solution sketch, I have to conclude the author of the solution sketch is quite imprecise in their writing.
The high level problem is to calculate an expected lifetime if components are chosen randomly by fair coin toss. This is what's leading to computing the LCM of all subsets -- all subsets effectively represent the sample space. You could end up with any possible set of components. The failure time for the device is based on the LCM of the set. The expected lifetime is therefore the average of the LCM of all sets.
Note that this ought to include the LCM of sets with only one item (in which case we'd assume the LCM to be the element itself). The solution sketch seems to sabotage, perhaps because they handled it in a less elegant manner.
What is meant by these states?
The sketch author only uses the word state twice, but apparently manages to switch meanings. In the first use of the word state it appears they're talking about a possible selection of components. In the second use they're likely talking about possible failure times. They could be muddling this terminology because their dynamic programming solution initializes values from one use of the word and the recurrence relation stems from the other.
Does dp stand for dynamic programming?
I would say either it does or it's a coincidence as the solution sketch seems to heavily imply dynamic programming.
If so, what recurrence relation is being solved? How is dp[i] computed from dp[i-1]?
All I can think is that in their solution, state i represents a time to failure , T(i), with the number of times this time to failure has been counted, dp[i]. The resulting sum would be to sum all dp[i] * T(i).
dp[i][0] would then be the failure times counted for only the first component. dp[i][1] would then be the failure times counted for the first and second component. dp[i][2] would be for the first, second, and third. Etc..
Initialize dp[i][0] with zeroes except for dp[T(c)][0] (where c is the first component considered) which should be 1 (since this component's failure time has been counted once so far).
To populate dp[i][n] from dp[i][n-1] for each component c:
For each i, copy dp[i][n-1] into dp[i][n].
Add 1 to dp[T(c)][n].
For each i, add dp[i][n-1] to dp[LCM(T(i), T(c))][n].
What is this doing? Suppose you knew that you had a time to failure of j, but you added a component with a time to failure of k. Regardless of what components you had before, your new time to fail is LCM(j, k). This follows from the fact that for two sets A and B, LCM(A union B} = LCM(LCM(A), LCM(B)).
Similarly, if we're considering a time to failure of T(i) and our new component's time to failure of T(c), the resultant time to failure is LCM(T(i), T(c)). Note that we recorded this time to failure for dp[i][n-1] configurations, so we should record that many new times to failure once the new component is introduced.
Why do the big primes not contribute to the number of states?
Each of them occurs either 0 or 1 times. Should the number of states not be multiplied by 2 for each of these primes (leading to a non-feasible state space again)?
You're right, of course. However, the solution sketch states that numbers with large primes are handled in another (unspecified) fashion.
What would happen if we did include them? The number of states we would need to represent would explode into an impractical number. Hence the author accounts for such numbers differently. Note that if a number less than or equal to 500 includes a prime larger than 19 the other factors multiply to 21 or less. This makes such numbers amenable for brute forcing, no tables necessary.
The first part of the editorial seems useful, but the second part is rather vague (and perhaps unhelpful; I'd rather finish this answer than figure it out).
Let's suppose for the moment that the input consists of pairwise distinct primes, e.g., 2, 3, 5, and 7. Then the answer (for summing all sets, where the LCM of 0 integers is 1) is
(1 + 2) (1 + 3) (1 + 5) (1 + 7),
because the LCM of a subset is exactly equal to the product here, so just multiply it out.
Let's relax the restriction that the primes be pairwise distinct. If we have an input like 2, 2, 3, 3, 3, and 5, then the multiplication looks like
(1 + (2^2 - 1) 2) (1 + (2^3 - 1) 3) (1 + (2^1 - 1) 5),
because 2 appears with multiplicity 2, and 3 appears with multiplicity 3, and 5 appears with multiplicity 1. With respect to, e.g., just the set of 3s, there are 2^3 - 1 ways to choose a subset that includes a 3, and 1 way to choose the empty set.
Call a prime small if it's 19 or less and large otherwise. Note that integers 500 or less are divisible by at most one large prime (with multiplicity). The small primes are more problematic. What we're going to do is to compute, for each possible small portion of the prime factorization of the LCM (i.e., one of the ~70,000 states), the sum of LCMs for the problem derived by discarding the integers that could not divide such an LCM and leaving only the large prime factor (or 1) for the other integers.
For example, if the input is 2, 30, 41, 46, and 51, and the state is 2, then we retain 2 as 1, discard 30 (= 2 * 3 * 5; 3 and 5 are small), retain 41 as 41 (41 is large), retain 46 as 23 (= 2 * 23; 23 is large), and discard 51 (= 3 * 17; 3 and 17 are small). Now, we compute the sum of LCMs using the previously described technique. Use inclusion-exclusion to get rid of the subsets whose LCM whose small portion properly divides the state instead of being exactly equal. Maybe I'll work a complete example later.
What is meant by these states?
I think here, states refer to if the number is in set B = {b0, b1, ..., bk-1} of LCMs of set A.
Does dp stand for dynamic programming and if so, what recurrence relation is being solved?
dp in the solution sketch stands for dynamic programming, I believe.
How is dp[i] computed from dp[i-1]?
It's feasible that we can figure out the state of next group of LCMs from previous states. So, we only need array of 2, and toggle back and forth.
Why do the big primes not contribute to the number of states? Each of them occurs either 0 or 1 times. Should the number of states not be multiplied by 2 for each of these primes (leading to a non-feasible state space again)?
We can use Prime Factorization and exponents only to present the number.
Here is one example.
6 = (2^1)(3^1)(5^0) -> state "1 1 0" to represent 6
18 = (2^1)(3^2)(5^0) -> state "1 2 0" to represent 18
Here is how we can get LMC of 6 and 18 using Prime Factorization
LCM (6,18) = (2^(max(1,1)) (3^ (max(1,2)) (5^max(0,0)) = (2^1)(3^2)(5^0) = 18
2^9 > 500, 3^6 > 500, 5^4 > 500, 7^4>500, 11^3 > 500, 13^3 > 500, 17^3 > 500, 19^3 > 500
we can use only count of exponents of prime number 2,3,5,7,11,13,17,19 to represent the LCMs in the set B = {b0, b1, ..., bk-1}
for the given set A = {a0, a1, ..., aN-1} (1 ≤ N ≤ 100), with 2 ≤ ai ≤ 500.
9 * 6 * 4 * 4 * 3 * 3 * 3 * 3 <= 70000, so we only need two of dp[9][6][4][4][3][3][3][3] to keep tracks of all LCMs' states. So, dp[70000][2] is enough.
I put together a small C++ program to illustrate how we can get sum of LCMs of the given set A = {a0, a1, ..., aN-1} (1 ≤ N ≤ 100), with 2 ≤ ai ≤ 500. In the solution sketch, we need to loop through 70000 max possible of LCMs.
int gcd(int a, int b) {
int remainder = 0;
do {
remainder = a % b;
a = b;
b = remainder;
} while (b != 0);
return a;
}
int lcm(int a, int b) {
if (a == 0 || b == 0) {
return 0;
}
return (a * b) / gcd(a, b);
}
int sum_of_lcm(int A[], int N) {
// get the max LCM from the array
int max = A[0];
for (int i = 1; i < N; i++) {
max = lcm(max, A[i]);
}
max++;
//
int dp[max][2];
memset(dp, 0, sizeof(dp));
int pri = 0;
int cur = 1;
// loop through n x 70000
for (int i = 0; i < N; i++) {
for (int v = 1; v < max; v++) {
int x = A[i];
if (dp[v][pri] > 0) {
x = lcm(A[i], v);
dp[v][cur] = (dp[v][cur] == 0) ? dp[v][pri] : dp[v][cur];
if ( x % A[i] != 0 ) {
dp[x][cur] += dp[v][pri] + dp[A[i]][pri];
} else {
dp[x][cur] += ( x==v ) ? ( dp[v][pri] + dp[v][pri] ) : ( dp[v][pri] ) ;
}
}
}
dp[A[i]][cur]++;
pri = cur;
cur = (pri + 1) % 2;
}
for (int i = 0; i < N; i++) {
dp[A[i]][pri] -= 1;
}
long total = 0;
for (int j = 0; j < max; j++) {
if (dp[j][pri] > 0) {
total += dp[j][pri] * j;
}
}
cout << "total:" << total << endl;
return total;
}
int test() {
int a[] = {2, 6, 7 };
int n = sizeof(a)/sizeof(a[0]);
int total = sum_of_lcm(a, n);
return 0;
}
Output
total:104
The states are one more than the powers of primes. You have numbers up to 2^8, so the power of 2 is in [0..8], which is 9 states. Similarly for the other states.
"dp" could well stand for dynamic programming, I'm not sure.
The recurrence relation is the heart of the problem, so you will learn more by solving it yourself. Start with some small, simple examples.
For the large primes, try solving a reduced problem without using them (or their equivalents) and then add them back in to see their effect on the final result.