In some conditions like Segment Tree or Binary Search, I have to get the average value of 2 number.
An easy solution is mid = (a + b) / 2. But when it comes to a and b have the same plus-minus signs, The a + b thing will overflow. For example, 1234567 + 2147483647, -1234567 + (-2147483647) in int32 type.
I searched online and got to know mid = (b - a) / 2 + a, it can avoid the situation above, but still not perfect. When it comes to a and b have the different plus-minus signs, mid = (a + b) / 2 won't overflow but (b - a) / 2 + a will. For example, -2147483648 + 2147483647 in int32 type.
To thoroughly solve this problem, I've written the code in pic below. I divide the 2 situations by the plus-minus signs. (I use some bit operations to improve its efficiency.) But it is way too complex for such a simple problem, right?
Is there some elegant solution to this problem?
I tried to divide the problem to 2 situations and solve them respectively.
But I still want a more elegant solution.
Got the answer!
mid = (a & b) + ((a ^ b) >> 1)
a & b keeps the same bits that a and b have in common, they need not to be distributed averagely. Like when you find the average value of 102 and 110, you don't need to calculate the 100 they have in common. You can just keep that, and deal with the 2 and 10 part, distribute them averagely to 2 number. As (102 + 110) / 2 = (2 * 100 + 2 + 10) / 2 = 100 + (2 + 10) / 2 = 100 + 6 = 106.
(a ^ b) >> 1 deals with the "2 and 10 part", it gets all the bits that a and b don't have in common, and divide it by 2.
Adds up 2 parts above so we get the average value of a and b. Not a strict proof, though.
Related
I am trying to fit 3 numbers inside 1 number.But numbers will be only between 0 and 11.So their (base) is 12.For example i have 7,5,2 numbers.I come up with something like this:
Three numbers into One number :
7x12=84
84x5=420
420+2=422
Now getting back Three numbers from One number :
422 MOD 12 = 2 (the third number)
422 - 2 = 420
420 / 12 = 35
And i understanded that 35 is multiplication of first and the second number (i.e 7 and 5)
And now i cant get that 7 and 5 anyone knows how could i ???
(I started typing this answer before the other one got posted, but this one is more specific to Arduino then the other one, so I'm leaving it)
The code
You can use bit shifting to get multiple small numbers into one big number, in code it would look like this:
int a, b, c;
//putting then together
int big = (a << 8) + (b << 4) + c;
//separating them again
a = (big >> 8) & 15;
b = (big >> 4) & 15;
c = big & 15;
This code only works when a, b and c are all in the range [0, 15] witch appears to be enough for you case.
How it works
The >> and << operators are the bitshift operators, in short a << n shifts every bit in a by n places to the left, this is equivalent to multiplying by 2^n. Similarly, a >> n shifts to to the right. An example:
11 << 3 == 120 //0000 1011 -> 0101 1000
The & operator performs a bitwise and on the two operands:
6 & 5 == 4 // 0110
// & 0101
//-> 0100
These two operators are combined to "pack" and "unpack" the three numbers. For the packing every small number is shifted a bit to the left and they are all added together. This is how the bits of big now look (there are 16 of them because ints in Arduino are 16 bits wide):
0000aaaabbbbcccc
When unpacking, the bits are shifted to the right again, and they are bitwise anded together with 15 to filter out any excess bits. This is what that last operation looks like to get b out again:
00000000aaaabbbb //big shifted 4 bits to the right
& 0000000000001111 //anded together with 15
-> 000000000000bbbb //gives the original number b
All is working exactly like in base 10 (or 16). Here after your corrected example.
Three numbers into One number :
7x12^2=1008
5*12^1=60
2*12^0=2
1008+60+2=1070
Now getting back Three numbers from One number :
1070 MOD 12 = 2 (the third number)
1070/12 = 89 (integer division) => 89 MOD 12 = 5
89 / 12 = 7
Note also that the maximum value will be 11*12*12+11*12+11=1727.
If this is really programming related, you will be using 16bits instead of 3*8 bits so sparing one byte. An easyer method not using base 12 would be fit each number into half a byte (better code efficiency and same transmission length):
7<<(4+4) + 5<<4 + 2 = 1874
1874 & 0x000F = 2
1874>>4 & 0x000F = 5
1874>>8 & 0x0F = 7
Because MOD(12) and division by 12 is much less efficient than working with powers of 2
you can use the principle of the positional notation to change from one or the other in any base
Treat yours numbers (n0,n1,...,nm) as a digit of a big number in the base B of your choosing so the new number is
N = n0*B^0 + n1*B^1 + ... + nm*B^m
to revert the process is also simple, while your number is greater than 0 find its modulo in respect to the base to get to get the first digit, then subtracts that digit and divide for the base, repeat until finish while saving each digit along the way
digit_list = []
while N > 0 do:
d = N mod B
N = (N - d) / B
digit_list.append( d )
then if N is N = n0*B^0 + n1*B^1 + ... + nm*B^m doing N mod B give you n0, then subtract it leaving you with n1*B^1 + ... + nm*B^m and divide by B to reduce the exponents of all B and that is the new N, N = n1*B^0 + ... + nm*B^(m-1) repetition of that give you all the digit you start with
here is a working example in python
def compact_num( num_list, base=12 ):
return sum( n*pow(base,i) for i,n in enumerate(num_list) )
def decompact_num( n, base=12):
if n==0:
return [0]
result = []
while n:
n,d = divmod(n,base)
result.append(d)
return result
example
>>> compact_num([2,5,7])
1070
>>> decompact_num(1070)
[2, 5, 7]
>>> compact_num([10,2],16)
42
>>> decompact_num(42,16)
[10, 2]
>>>
Suppose we have a set like {1,2,3} then there is only one way to choose 3 consecutive numbers... it's the set {1,2,3}...
For a set of {1,2,3,4} we have 3 ways: 123 234 1234
(technically these are unordered sets of numbers, but writing them consecutively helps)
f(5) ; {1,2,3,4,5} -> 8 ways: 123 1234 1235 12345 234 2345 345 1345
f(6) ; {1,2,3,4,5,6} -> 20 ways: ...
f(7) ; {1,2,3,4,5,6,7} -> 47 ways: ...
So for a given N, I can get the answer by applying brute force, and calculating all such subset having 3 or more consecutive number.
Here I am just trying to find out a pattern, a technique to get the number of all such subset for a given N.
The problem is further generalized to .....discover m consecutive number within a set of size N.
There is a bijection between this problem and "the number of N-digit binary numbers with at least three consecutive 1s in a row somewhere" (the bijection being a number is 0 if excluded in the subset, and 1 if included in the subset).
This is a known problem, and should be enough information to google for a result, if you search for number of n-digit binary strings with m consecutive 1s, the second hit is Finding all n digit binary numbers with r adjacent digits as 1
Alternatively you can just look it up as http://oeis.org/search?q=0%2C0%2C1%2C3%2C8%2C20%2C47 (based on the brute-forcing you did for the first few terms) - resulting in an explicit formula of 2^n - tribonacci(n+3), see here for an explicit formula for tribonacci numbers. It also gives a recurrence relation. The analogy given is "probability (out of 2^n) of getting at least 1 run of 3 heads within n flips of a fair coin"
I can only assume that the answer to the general problem is 2^n - Fm(n+m), where Fm is the mth n-step Fibonacci number (edit: that does seem to be the case)
This sounds like homework to me, so I'll just get you started. FoOne approach is to think of the Lowest and Highest members of the run, L and H. If the set size is N and your minimum run length is M, then for each possible position P of L, you can work out how many positions of H there are....
With a bit of python code, we can investigate this:
y = set()
def cons(li, num):
if len(li) < num:
return
if len(li) == num:
y.add(tuple([i for i in li]))
else:
y.add(tuple([i for i in li]))
cons(li[1:], num)
cons(li[:-1], num)
This solution will be quite slow (it's exponential in complexity, actually), but try it out for a few small list sizes and I think you should be able to pick up the pattern.
Not sure if you mean consecutive or not. If not, then for {1, 2, 3, 4} there are 4 possibilities: {1, 2, 3} {2, 3, 4} {1, 3, 4} {1, 2, 3, 4}
I think you can calculate the solution with N!/3! where N! = N*(N-1)(N-2)...*1.
Quick answer:
Sequences(n) = (n-1)*(n-2) / 2
Long answer:
You can do this by induction. First, I'm going to re-state the problem, because your problem statement isn't clear enough.
Rule 1: For all sets of consecutive numbers 1..n where n is 2 or more
Rule 2: Count the subsets S(n) of consecutive numbers m..m+q where q is 2 or more
S(n=3)
By inspection we find only one - 123
S(n=4)
By inspection we find 3! - 123 234 and 1234
Note that S(4) contains S(3), plus two new ones... both include the new digit 4... hmm.
S(n=5)
By inspection we find ... S(n=4) as well as 345 2345 and 12345. That's 3+3=6 total.
I think there's a pattern forming here. Let's define a new function T.
Rule 3: S(n) = S(n-1) + T(n) ... for some T.
We know that S(n) contains the digit n, and should have spotted by now that S(n) also contains (as a subcomponent) all sequences of length 3 to n that include the digit n. We know they cannot be in S(n-1) so they must be in T(n).
Rule 4: T(n) contains all sequence ending in n that are of length 3 to n.
How many sequences are in S(n)?
Let's look back at S(3) S(4) and S(5), and incorporate T(n):
S(3) = S(3)
S(4) = S(3) + T(4)
S(5) = S(4) + T(5) = S(3) + T(4) + T(5)
let's generalise:
S(n) = S(3) + T(f) for all f from 4 to n.
So how many are in a given T?
Look back at rule 5 - how many sequences does it describe?
For T(4) it describes all sequences 3 and longer ending in 4. (that's 234)
For T(5) it describes all sequences 3 and longer ending in 5. (that's 345 2345 = 2)
T count Examples
4 2 1234 234
5 3 12345 2345 345
6 4 123456 23456 3456 456
Looks awfully like T(n) is simply n-2!
So
S(6) = T(6) + T(5) + T(4) + S(3)
10 = 4 + 3 + 2 + 1
And
S(7) = 15 = 5 + 4 + 3 + 2 + 1
S(8) = 21 = 6 + 5 + 4 + 3 + 2 + 1
Turning this into a formula
What's 2 * S(8)?
42 = 6 + 5 + 4 + 3 + 2 + 1 + 1 + 2 + 3 + 4 + 5 + 6
Add each pair of biggest and smallest numbers:
42 = 7 + 7 + 7 + 7 + 7 + 7
42 = 7 * 6
But that's 2 * S(8), so
S(8) = 42/2 = 21 = 7 * 6 / 2
This generalizes:
S(n) = (n-1)*(n-2) / 2
Let's check this works:
S(3) = 2*1/2 = 1
S(4) = 3*2/2 = 3
S(5) = 4*3/2 = 6
S(6) = 5*4/2 = 10
I'm satisfied.
I would like to compute the average of three numbers, like:
d = int(round((a + b + c) / 3.0))
Where a, b, c, d are integers.
Is it possible to get the same result using just integers?
I'm interested in this because of performance reasons, I assume doing the math using integers should be faster than using floats.
The example above converts the integers to floats, calculates the result, rounds it and converts back to integer. Is it possible to avoid the int <-> float conversions?
Given the requirements for 1, 1, 2 -> 1; 1, 2, 2 -> 2 then this can be done using integer division.
Using // for the integer division and n for number of elements.
average = ( a+ b + c + .... + n//2 ) // n
ie sum up all the values and then add a number to deal with rounding.
As noted in #Henrik's answer this assumes that all numbers are positive.
(a + b + c + 1) / 3
Explanation: if (a + b + c) % 3 == 1, it's rounded down; if (a + b + c) % 3 == 2, it's rounded up.
At least this should work for a + b + c >= 0. You may need to treat negative values separately.
A variant of Mahmoud's answer using only one division:
d = (((a+b+c) * 10) + 15) / 30
It should be as simple as that:
d = (((a+b+c) * 10) / 3 + 5) / 10;
I have this recursive function:
f(n) = 2 * f(n-1) + 3 * f(n-2) + 4
f(1) = 2
f(2) = 8
I know from experience that explicit form of it would be:
f(n) = 3 ^ n - 1 // pow(3, n) - 1
I wanna know if there's any way to prove that. I googled a bit, yet didn't find anything simple to understand. I already know that generation functions probably solve it, they're too complex, I'd rather not get into them. I'm looking for a simpler way.
P.S.
If it helps I remember something like this solved it:
f(n) = 2 * f(n-1) + 3 * f(n-2) + 4
// consider f(n) = x ^ n
x ^ n = 2 * x ^ (n-1) + 3 * x ^ (n-2) + 4
And then you somehow computed x that lead to explicit form of the recursive formula, yet I can't quite remember
f(n) = 2 * f(n-1) + 3 * f(n-2) + 4
f(n+1) = 2 * f(n) + 3 * f(n-1) + 4
f(n+1)-f(n) = 2 * f(n) - 2 * f(n-1) + 3 * f(n-1) - 3 * f(n-2)
f(n+1) = 3 * f(n) + f(n-1) - 3 * f(n-2)
Now the 4 is gone.
As you said the next step is letting f(n) = x ^ n
x^(n+1) = 3 * x^n + x^(n-1) - 3 * x^(n-2)
divide by x^(n-2)
x^3 = 3 * x^2 + x - 3
x^3 - 3 * x^2 - x + 3 = 0
factorise to find x
(x-3)(x-1)(x+1) = 0
x = -1 or 1 or 3
f(n) = A * (-1)^n + B * 1^n + C * 3^n
f(n) = A * (-1)^n + B + C * 3^n
Now find A,B and C using the values you have
f(1) = 2; f(2) = 8; f(3) = 26
f(1) = 2 = -A + B + 3C
f(2) = 8 = A + B + 9C
f(3) = 26 = -A + B + 27C
solving for A,B and C:
f(3)-f(1) = 24 = 24C => C = 1
f(2)-f(1) = 6 = 2A + 6 => A = 0
2 = B + 3 => B = -1
Finally
f(n) = 3^n - 1
Ok, I know you didn't want generating functions (GF from now on) and all the complicated stuff, but my problem turned out to be nonlinear and simple linear methods didn't seem to work. So after a full day of searching, I found the answer and hopefully these findings will be of help to others.
My problem: a[n+1]= a[n]/(1+a[n]) (i.e. not linear (nor polynomial), but also not completely nonlinear - it is a rational difference equation)
if your recurrence is linear (or polynomial), wikihow has step-by-step instructions (with and without GF)
if you want to read something about GF, go to this wiki, but I didn't get it till I started doing examples (see next)
GF usage example on Fibonacci
if the previous example didn't make sense, download GF book and read the simplest GF example (section 1.1, ie a[n+1]= 2 a[n]+1, then 1.2, a[n+1]= 2 a[n]+1, then 1.3 - Fibonacci)
(while I'm on the book topic) templatetypedef mentioned Concrete Mathematics, download here, but I don't know much about it except it has a recurrence, sums, and GF chapter (among others) and a table of simple GFs on page 335
as I dove deeper for nonlinear stuff, I saw this page, using which I failed at z-transforms approach and didn't try linear algebra, but the link to rational difference eqn was the best (see next step)
so as per this page, rational functions are nice because you can transform them into polynomials and use linear methods of step 1. 3. and 4. above, which I wrote out by hand and probably made some mistake, because (see 8)
Mathematica (or even the free WolframAlpha) has a recurrence solver, which with RSolve[{a[n + 1] == a[n]/(1 + a[n]), a[1] == A}, a[n], n] got me a simple {{a[n] -> A/(1 - A + A n)}}. So I guess I'll go back and look for mistake in hand-calculations (they are good for understanding how the whole conversion process works).
Anyways, hope this helps.
In general, there is no algorithm for converting a recursive form into an iterative one. This problem is undecidable. As an example, consider this recursive function definition, which defines the Collatz sequence:
f(1) = 0
f(2n) = 1 + f(n)
f(2n + 1) = 1 + f(6n + 4)
It's not known whether or not this is even a well-defined function or not. Were an algorithm to exist that could convert this into a closed-form, we could decide whether or not it was well-defined.
However, for many common cases, it is possible to convert a recursive definition into an iterative one. The excellent textbook Concrete Mathematics spends much of its pages showing how to do this. One common technique that works quite well when you have a guess of what the answer is is to use induction. As an example for your case, suppose that you believe that your recursive definition does indeed give 3^n - 1. To prove this, try proving that it holds true for the base cases, then show that this knowledge lets you generalize the solution upward. You didn't put a base case in your post, but I'm assuming that
f(0) = 0
f(1) = 2
Given this, let's see whether your hunch is correct. For the specific inputs of 0 and 1, you can verify by inspection that the function does compute 3^n - 1. For the inductive step, let's assume that for all n' < n that f(n) = 3^n - 1. Then we have that
f(n) = 2f(n - 1) + 3f(n - 2) + 4
= 2 * (3^{n-1} - 1) + 3 * (3^{n-2} - 1) + 4
= 2 * 3^{n-1} - 2 + 3^{n-1} - 3 + 4
= 3 * 3^{n-1} - 5 + 4
= 3^n - 1
So we have just proven that this recursive function does indeed produce 3^n - 1.
I am embarrassed to ask such question; but I haven't use math for a long time I can not recall many concepts learned many years ago.
In the url http://www.javadev.org/files/Ranking.pdf, an example is used for illustrate the page rank mechanism. The relation between page A, B, and C is A links to B and C, B links to C, and C links to A. So the PageRank equation is as below
Equation A)
PR(A) = 0.5 + 0.5 PR(C)
PR(B) = 0.5 + 0.5 (PR(A) / 2)
PR(C) = 0.5 + 0.5 (PR(A) / 2 + PR(B))
and it comes up with the result
Result B)
PR(A) = 14/13 = 1.07692308
PR(B) = 10/13 = 0.76923077
PR(C) = 15/13 = 1.15384615
My question is how Result B is derived from Equation A?
I try e.g. replacing PR(C) in equation PR(A)
PR(A) = 0.5 + 0.5 (0.5 + 0.5 (PR(A) / 2 + PR(B)))
this seems to end up with an infinite loop. So I am confused how it can derive the result e.g. PR(A) value is 1.07692308?
Appologize for such stupid question.
I appreciate any advice.
Not a stupid question, you're just rusty.
Take your equation and multiply by 16 (not really necessary, but it makes things look nicer):
16 PR(A) = 12 + 2 PR(A) + 4 PR(B)
Now subtract 2 PR(A) from both sides:
14 PR(A) = 12 + 4 PR(B)
Now replace PR(B), using the second part of "equation A":
14 PR(A) = 12 + 2 + PR(A)
13 PR(A) = 14
PR(A) = 14/13
And the others follow the same way. If you find that an equation winds up being the same on both sides (X = X), it probably means that you did the same substitution twice; just back up and try again. With a little practice you'll get the hang of it.