I have a function that generates binary sequences with a fixed number of 1's (the rest are 0's). I need a function that takes a sequences and returns the position of that sequence in lexicographic order. For example, the 10 sequences of length 5 with 3 1's are
0 0 1 1 1
0 1 0 1 1
0 1 1 0 1
0 1 1 1 0
1 0 0 1 1
1 0 1 0 1
1 0 1 1 0
1 1 0 0 1
1 1 0 1 0
1 1 1 0 0
I need a function that takes, for example 0 1 1 0 1 and returns 3 since it's the third in the list.
The only thing I can think of, which is way too inefficient, is to generate all of the sequences (easy), store them (takes too much space), then search for the given sequence in the list (too slow), and return its position. Is there a faster way to do this? Some easy trick that I don't see?
We call the set of sequences of length n with k 1's binseq(n,k). This problem can then be solved recursively, as follows:
Base case: If S has length 1, it's in position 1.
If S starts with a 0, its position is the same as the position of tail(S) (S with the first element removed) in binseq(n-1, k).
If S starts with a 1, its position is equal to the position of tail(S) in binseq(n-1, k-1) plus the number of sequences in binseq(n-1, k).
In python code:
#!/usr/bin/env python
def binom(n, k):
result = 1
for i in range(1, k+1):
result = result * (n-i+1) / i
return result
def lexpos(seq):
if len(seq) == 1:
return 1
elif seq[0] == 0:
return lexpos(seq[1:])
else:
return binom(len(seq)-1, seq.count(1)) + lexpos(seq[1:])
Or the iterative version, as suggested by Abhishek Bansal:
def lexpos_iter(seq):
pos = 1
for i in xrange(len(seq)):
if seq[i] == 1:
pos += binom(len(seq)-i-1, seq[i:].count(1))
return pos
Related
i seen a post on the site about it and i didn't understand the answer, can i get explanation please:
question:
Write code to determine if a number is divisible by 3. The input to the function is a single bit, 0 or 1, and the output should be 1 if the number received so far is the binary representation of a number divisible by 3, otherwise zero.
Examples:
input "0": (0) output 1
inputs "1,0,0": (4) output 0
inputs "1,1,0,0": (6) output 1
This is based on an interview question. I ask for a drawing of logic gates but since this is stackoverflow I'll accept any coding language. Bonus points for a hardware implementation (verilog etc).
Part a (easy): First input is the MSB.
Part b (a little harder): First input is the LSB.
Part c (difficult): Which one is faster and smaller, (a) or (b)? (Not theoretically in the Big-O sense, but practically faster/smaller.) Now take the slower/bigger one and make it as fast/small as the faster/smaller one.
answer:
State table for LSB:
S I S' O
0 0 0 1
0 1 1 0
1 0 2 0
1 1 0 1
2 0 1 0
2 1 2 0
Explanation: 0 is divisible by three. 0 << 1 + 0 = 0. Repeat using S = (S << 1 + I) % 3 and O = 1 if S == 0.
State table for MSB:
S I S' O
0 0 0 1
0 1 2 0
1 0 1 0
1 1 0 1
2 0 2 0
2 1 1 0
Explanation: 0 is divisible by three. 0 >> 1 + 0 = 0. Repeat using S = (S >> 1 + I) % 3 and O = 1 if S == 0.
S' is different from above, but O works the same, since S' is 0 for the same cases (00 and 11). Since O is the same in both cases, O_LSB = O_MSB, so to make MSB as short as LSB, or vice-versa, just use the shortest of both.
thanks for the answers in advanced.
Well, I suppose the question isn't entirely off-topic, since you asked about logic design, but you'll have to do the coding yourself.
You have 3 states in the S column. These track the value of the current full input mod 3. So, S0 means the current input mod 3 is 0, and so is divisible by 0 (remember also that 0 is divisible by 3). S1 means the remainder is 1, S2 means that the remainder is 2.
The I column gives the current input (0 or 1), and S' gives the next state (in other words, the new number mod 3).
For 'LSB', the new number is the old number << 1, plus either 0 or 1. Write out the table. For starters, if the old modulo was 0, then the new modulo will be 0 if the input bit was 0, and will be 1 if the new input was 1. This gives you the first 2 rows in the first table. Filling in the rest is left as an exercise for you.
Note that the O column is just 1 if the next state is 0, as expected.
I have a matrix, which contains N entries each with M rows. Each row contains of 0s and 1s. I want to create a second matrix with the same size, but in each row only one 1 should be left, every other value should be 0. Which value should be 1 should be chosen randomly.
E.g.:
0 1 1 0 1
1 1 0 0 1
0 0 1 1 0
->
0 1 0 0 0
1 0 0 0 0
0 0 0 1 0
Read the documentation of find and randperm
%//preallocate the output matrix
out = zeros(size(a));
%for each row, take a random sample from the indices holding value 1
for i = 1:size(a,1)
temp2 = find(a(i,:));
out(i,temp2(randperm(numel(temp2))(1))) = 1;
end
Watch the code in action here
I need an algorithm in Matlab which counts how many adjacent and non-overlapping (1,1) I have in each row of a matrix A mx(n*2) without using loops. E.g.
A=[1 1 1 0 1 1 0 0 0 1; 1 0 1 1 1 1 0 0 1 1] %m=2, n=5
Then I want
B=[2;3] %mx1
Specific case
Assuming A to have ones and zeros only, this could be one way -
B = sum(reshape(sum(reshape(A',2,[]))==2,size(A,2)/2,[]))
General case
If you are looking for a general approach that must work for all integers and a case where you can specify the pattern of numbers, you may use this -
patt = [0 1] %%// pattern to be found out
B = sum(reshape(ismember(reshape(A',2,[])',patt,'rows'),[],2))
Output
With patt = [1 1], B = [2 3]
With patt = [0 1], B = [1 0]
you can use transpose then reshape so each consecutive values will now be in a row, then compare the top and bottom row (boolean compare or compare the sum of each row to 2), then sum the result of the comparison and reshape the result to your liking.
in code, it would look like:
A=[1 1 1 0 1 1 0 0 0 1; 1 0 1 1 1 1 0 0 1 1] ;
m = size(A,1) ;
n = size(A,2)/2 ;
Atemp = reshape(A.' , 2 , [] , m ) ;
B = squeeze(sum(sum(Atemp)==2))
You could pack everything in one line of code if you want, but several lines is usually easier for comprehension. For clarity, the Atemp matrix looks like that:
Atemp(:,:,1) =
1 1 1 0 0
1 0 1 0 1
Atemp(:,:,2) =
1 1 1 0 1
0 1 1 0 1
You'll notice that each row of the original A matrix has been broken down in 2 rows element-wise. The second line will simply compare the sum of each row with 2, then sum the valid result of the comparisons.
The squeeze command is only to remove the singleton dimensions not necessary anymore.
you can use imresize , for example
imresize(A,[size(A,1),size(A,2)/2])>0.8
ans =
1 0 1 0 0
0 1 1 0 1
this places 1 where you have [1 1] pairs... then you can just use sum
For any pair type [x y] you can :
x=0; y=1;
R(size(A,1),size(A,2)/2)=0; % prealocarting memory
for n=1:size(A,1)
b=[A(n,1:2:end)' A(n,2:2:end)']
try
R(n,find(b(:,1)==x & b(:,2)==y))=1;
end
end
R =
0 0 0 0 1
0 0 0 0 0
With diff (to detect start and end of each run of ones) and accumarray (to group runs of the same row; each run contributes half its length rounded down):
B = diff([zeros(1,size(A,1)); A.'; zeros(1,size(A,1))]); %'// columnwise is easier
[is js] = find(B==1); %// rows and columns of starts of runs of ones
[ie je] = find(B==-1); %// rows and columns of ends of runs of ones
result = accumarray(js, floor((ie-is)/2)); %// sum values for each row of A
Is it possible to decompose a matrix A having n rows and n columns to sum of m [n x n] permutation matrices. where m is the number of 1's in each row and each column in matrix A?
UPDATE:
yes, this is possible. I came across such an exmaple which is shown below - but How can we generalize the answer?
What you want is called a 1-factorization. One algorithm is repeatedly to find a perfect matching and remove it; probably there are others.
For the first permutation matrix, take the first 1 in the first row. For the second row, take the first 1 that is in a column you don't already have. For the third row, take the first 1 that is in a column you don't already have. And so on. Do this for all rows.
You now have one permutation matrix.
Next subtract your first permutation matrix from the original. This new matrix now has m-1 ones in each row and column. So repeat the process m-1 more times, and you'll have your m permutation matrices.
You can skip the last step, because a matrix with one 1 in each row and column already is a permutation matrix. There's no need to do any calculations.
This is a greedy algorithm that doesn't always work. We can make it work by changing the selection rule slightly. See below:
For your example:
1 0 1 1
A = 1 1 0 1
1 1 1 0
0 1 1 1
In the first step, we pick (1,1) for the first row, (2,2) for the second row, (3,3) for the thrid row and (4,4) for the 4th row. We then have:
1 0 0 0 0 0 1 1
A = 0 1 0 0 + 1 0 0 1
0 0 1 0 1 1 0 0
0 0 0 1 0 1 1 0
The first matrix is a permutation matrix. The second matrix has exactly two 1's in each row and column. So we pick, in order: (1,3), (2,1), (3,2) and... we're in trouble: the rows that contain a 1 in column 4 have already been used.
So how do we fix this? Well, we can keep track of the number of 1's remaining in each column. Instead of picking the first column that is unused, we pick the column with the lowest number of 1's remaining. For the second matrix above:
0 0 1 1 0 0 X 0 0 0 X 0 0 0 X 0
B = 1 0 0 1 --> 1 0 0 1 --> 0 0 0 X --> 0 0 0 X
1 1 0 0 1 1 0 0 1 1 0 0 X 0 0 0
0 1 1 0 0 1 1 0 0 1 1 0 0 1 1 0
------- ------- ------- -------
2 2 2 2 2 2 X 1 1 2 X X X 1 X X
So we would pick column 4 in the second step, column 1 in the 3rd step, and column 2 in the 4th step.
There can always be only one column with one remaining 1. The other 1's must have been taken away in m-1 previous rows. If you had two such columns, one of them would have had to have been picked as the minimum column before.
This can be done easily using a recursive (backtracking OR depth-first traversal) algorithm. Here is the pseudo-code for its solution:
void printPermutationMatrices(const int OrigMat[][], int permutMat[], int curRow, const int n){
//curPermutMatrix is 1-D array where value of ith element contains the value of column where 1 is placed in ith row
if(curRow == n){//Base case
//do stuff with permutMat[]
printPermutMat(permutMat);
return;
}
for(int col=0; col<n; col++){//try to place 1 in cur_row in each col if possible and go further to next row in recursion
if(origM[cur_row][col] == 1){
permutMat[cur_row] = col;//choose this col for cur_row
if there is no conflict to place a 1 in [cur_row, col] in permutMat[]
perform(origM, curPermutMat, curRow+1, n);
}
}
}
Here is how to call from your main function:
int[] permutMat = new int[n];
printPermutationMatrices(originalMatrix, permutMat, 0, n);
I'm trying to come up with an algorithm that will print out all possible ways to sum N integers so that they total a given value.
Example. Print all ways to sum 4 integers so that they sum up to be 5.
Result should be something like:
5 0 0 0
4 1 0 0
3 2 0 0
3 1 1 0
2 3 0 0
2 2 1 0
2 1 2 0
2 1 1 1
1 4 0 0
1 3 1 0
1 2 2 0
1 2 1 1
1 1 3 0
1 1 2 1
1 1 1 2
This is based off Alinium's code.
I modified it so it prints out all the possible combinations, since his already does all the permutations.
Also, I don't think you need the for loop when n=1, because in that case, only one number should cause the sum to equal value.
Various other modifications to get boundary cases to work.
def sum(n, value):
arr = [0]*n # create an array of size n, filled with zeroes
sumRecursive(n, value, 0, n, arr);
def sumRecursive(n, value, sumSoFar, topLevel, arr):
if n == 1:
if sumSoFar <= value:
#Make sure it's in ascending order (or only level)
if topLevel == 1 or (value - sumSoFar >= arr[-2]):
arr[(-1)] = value - sumSoFar #put it in the n_th last index of arr
print arr
elif n > 0:
#Make sure it's in ascending order
start = 0
if (n != topLevel):
start = arr[(-1*n)-1] #the value before this element
for i in range(start, value+1): # i = start...value
arr[(-1*n)] = i # put i in the n_th last index of arr
sumRecursive(n-1, value, sumSoFar + i, topLevel, arr)
Runing sums(4, 5) returns:
[0, 0, 0, 5]
[0, 0, 1, 4]
[0, 0, 2, 3]
[0, 1, 1, 3]
[1, 1, 1, 2]
In pure math, a way of summing integers to get a given total is called a partition. There is a lot of information around if you google for "integer partition". You are looking for integer partitions where there are a specific number of elements. I'm sure you could take one of the known generating mechanisms and adapt for this extra condition. Wikipedia has a good overview of the topic Partition_(number_theory). Mathematica even has a function to do what you want: IntegerPartitions[5, 4].
The key to solving the problem is recursion. Here's a working implementation in python. It prints out all possible permutations that sum up to the total. You'll probably want to get rid of the duplicate combinations, possibly by using some Set or hashing mechanism to filter them out.
def sum(n, value):
arr = [0]*n # create an array of size n, filled with zeroes
sumRecursive(n, value, 0, n, arr);
def sumRecursive(n, value, sumSoFar, topLevel, arr):
if n == 1:
if sumSoFar > value:
return False
else:
for i in range(value+1): # i = 0...value
if (sumSoFar + i) == value:
arr[(-1*n)] = i # put i in the n_th last index of arr
print arr;
return True
else:
for i in range(value+1): # i = 0...value
arr[(-1*n)] = i # put i in the n_th last index of arr
if sumRecursive(n-1, value, sumSoFar + i, topLevel, arr):
if (n == topLevel):
print "\n"
With some extra effort, this can probably be simplified to get rid of some of the parameters I am passing to the recursive function. As suggested by redcayuga's pseudo code, using a stack, instead of manually managing an array, would be a better idea too.
I haven't tested this:
procedure allSum (int tot, int n, int desiredTotal) return int
if n > 0
int i =
for (int i = tot; i>=0; i--) {
push i onto stack;
allSum(tot-i, n-1, desiredTotal);
pop top of stack
}
else if n==0
if stack sums to desiredTotal then print the stack end if
end if
I'm sure there's a better way to do this.
i've find a ruby way with domain specification based on Alinium's code
class Domain_partition
attr_reader :results,
:domain,
:sum,
:size
def initialize(_dom, _size, _sum)
_dom.is_a?(Array) ? #domain=_dom.sort : #domain= _dom.to_a
#results, #sum, #size = [], _sum, _size
arr = [0]*size # create an array of size n, filled with zeroes
sumRecursive(size, 0, arr)
end
def sumRecursive(n, sumSoFar, arr)
if n == 1
#Make sure it's in ascending order (or only level)
if sum - sumSoFar >= arr[-2] and #domain.include?(sum - sumSoFar)
final_arr=Array.new(arr)
final_arr[(-1)] = sum - sumSoFar #put it in the n_th last index of arr
#results<<final_arr
end
elsif n > 1
#********* dom_selector ********
n != size ? start = arr[(-1*n)-1] : start = domain[0]
dom_bounds=(start*(n-1)..domain.last*(n-1))
restricted_dom=domain.select do |x|
if x < start
false; next
end
if size-n > 0
if dom_bounds.cover? sum-(arr.first(size-n).inject(:+)+x) then true
else false end
else
dom_bounds.cover?(sum+x) ? true : false
end
end # ***************************
for i in restricted_dom
_arr=Array.new(arr)
_arr[(-1*n)] = i
sumRecursive(n-1, sumSoFar + i, _arr)
end
end
end
end
a=Domain_partition.new (-6..6),10,0
p a
b=Domain_partition.new [-4,-2,-1,1,2,3],10,0
p b
If you're interested in generating (lexically) ordered integer partitions, i.e. unique unordered sets of S positive integers (no 0's) that sum to N, then try the following. (unordered simply means that [1,2,1] and [1,1,2] are the same partition)
The problem doesn't need recursion and is quickly handled because the concept of finding the next lexical restricted partition is actually very simple...
In concept: Starting from the last addend (integer), find the first instance where the difference between two addends is greater than 1. Split the partition in two at that point. Remove 1 from the higher integer (which will be the last integer in one part) and add 1 to the lower integer (the first integer of the latter part). Then find the first lexically ordered partition for the latter part having the new largest integer as the maximum addend value. I use Sage to find the first lexical partition because it's lightening fast, but it's easily done without it. Finally, join the two portions and voila! You have the next lexical partition of N having S parts.
e.g. [6,5,3,2,2] -> [6,5],[3,2,2] -> [6,4],[4,2,2] -> [6,4],[4,3,1] -> [6,4,4,3,1]
So, in Python and calling Sage for the minor task of finding the first lexical partition given n and s parts...
from sage.all import *
def most_even_partition(n,s): # The main function will need to recognize the most even partition possible (i.e. last lexical partition) so it can loop back to the first lexical partition if need be
most_even = [int(floor(float(n)/float(s)))]*s
_remainder = int(n%s)
j = 0
while _remainder > 0:
most_even[j] += 1
_remainder -= 1
j += 1
return most_even
def portion(alist, indices):
return [alist[i:j] for i, j in zip([0]+indices, indices+[None])]
def next_restricted_part(p,n,s):
if p == most_even_partition(n,s):return Partitions(n,length=s).first()
for i in enumerate(reversed(p)):
if i[1] - p[-1] > 1:
if i[0] == (s-1):
return Partitions(n,length=s,max_part=(i[1]-1)).first()
else:
parts = portion(p,[s-i[0]-1]) # split p (soup?)
h1 = parts[0]
h2 = parts[1]
next = list(Partitions(sum(h2),length=len(h2),max_part=(h2[0]-1)).first())
return h1+next
If you want zeros (not actual integer partitions), then the functions only need small modifications.
Try this code. I hope it is easier to understand. I tested it, it generate correct sequence.
void partition(int n, int m = 0)
{
int i;
// if the partition is done
if(n == 0){
// Output the result
for(i = 0; i < m; ++i)
printf("%d ", list[i]);
printf("\n");
return;
}
// Do the split from large to small int
for(i = n; i > 0; --i){
// if the number not partitioned or
// willbe partitioned no larger than
// previous partition number
if(m == 0 || i <= list[m - 1]){
// store the partition int
list[m] = i;
// partition the rest
partition(n - i, m + 1);
}
}
}
Ask for clarification, if required.
The is One of the output
6
5 1
4 2
4 1 1
3 3
3 2 1
3 1 1 1
2 2 2
2 2 1 1
2 1 1 1 1
1 1 1 1 1 1
10
9 1
8 2
8 1 1
7 3
7 2 1
7 1 1 1
6 4
6 3 1
6 2 2
6 2 1 1
6 1 1 1 1
5 5
5 4 1
5 3 2
5 3 1 1
5 2 2 1
5 2 1 1 1
5 1 1 1 1 1
4 4 2
4 4 1 1
4 3 3
4 3 2 1
4 3 1 1 1
4 2 2 2
4 2 2 1 1
4 2 1 1 1 1
4 1 1 1 1 1 1
3 3 3 1
3 3 2 2
3 3 2 1 1
3 3 1 1 1 1
3 2 2 2 1
3 2 2 1 1 1
3 2 1 1 1 1 1
3 1 1 1 1 1 1 1
2 2 2 2 2
2 2 2 2 1 1
2 2 2 1 1 1 1
2 2 1 1 1 1 1 1
2 1 1 1 1 1 1 1 1
1 1 1 1 1 1 1 1 1 1