Given a list of numbers, say [4 5 2 3], I need to maximize the sum obtained according to the following set of rules:
I need to select a number from the list and that number will be removed.
Eg. selecting 2 will have the list as [4 5 3].
If the number to be removed has two neighbours then I should get the result of this selection as the product of the currently selected number with one of its neighbours and this product summed up with the other neighbour. eg.: if I select 2 then I can have the result of this selction as 2 * 5 + 3.
If I select a number with only one neighbour then the result is the product of the selected number with its neighbour.
When their is only one number left then it is just added to the result till now.
Following these rules, I need to select the numbers in such an order that the result is maximized.
For the above list, if the order of selction is 4->2->3->5 then the sum obtained is 53 which is the maximum.
I am including a program which lets you pass as input the set of elements and gives all possible sums and also indicates the max sum.
Here's a link.
import itertools
l = [int(i) for i in input().split()]
p = itertools.permutations(l)
c, cs = 1, -1
mm = -1
for i in p:
var, s = l[:], 0
print(c, ':', i)
c += 1
for j in i:
print(' removing: ', j)
pos = var.index(j)
if pos == 0 or pos == len(var) - 1:
if pos == 0 and len(var) != 1:
s += var[pos] * var[pos + 1]
var.remove(j)
elif pos == 0 and len(var) == 1:
s += var[pos]
var.remove(j)
if pos == len(var) - 1 and pos != 0:
s += var[pos] * var[pos - 1]
var.remove(j)
else:
mx = max(var[pos - 1], var[pos + 1])
mn = min(var[pos - 1], var[pos + 1])
s += var[pos] * mx + mn
var.remove(j)
if s > mm:
mm = s
cs = c - 1
print(' modified list: ', var, '\n sum:', s)
print('MAX SUM was', mm, ' at', cs)
Consider 4 variants of the problem: those where every element gets consumed, and those where either the left, the right, or both the right and left elements are not consumed.
In each case, you can consider the last element to be removed, and this breaks the problem down into 1 or 2 subproblems.
This solves the problem in O(n^3) time. Here's a python program that solves the problem. The 4 variants of solve_ correspond to none, one or the other, or both of the endpoints being fixed. No doubt this program can be reduced (there's a lot of duplication).
def solve_00(seq, n, m, cache):
key = ('00', n, m)
if key in cache:
return cache[key]
assert m >= n
if n == m:
return seq[n]
best = -1e9
for i in range(n, m+1):
left = solve_01(seq, n, i, cache) if i > n else 0
right = solve_10(seq, i, m, cache) if i < m else 0
best = max(best, left + right + seq[i])
cache[key] = best
return best
def solve_01(seq, n, m, cache):
key = ('01', n, m)
if key in cache:
return cache[key]
assert m >= n + 1
if m == n + 1:
return seq[n] * seq[m]
best = -1e9
for i in range(n, m):
left = solve_01(seq, n, i, cache) if i > n else 0
right = solve_11(seq, i, m, cache) if i < m - 1 else 0
best = max(best, left + right + seq[i] * seq[m])
cache[key] = best
return best
def solve_10(seq, n, m, cache):
key = ('10', n, m)
if key in cache:
return cache[key]
assert m >= n + 1
if m == n + 1:
return seq[n] * seq[m]
best = -1e9
for i in range(n+1, m+1):
left = solve_11(seq, n, i, cache) if i > n + 1 else 0
right = solve_10(seq, i, m, cache) if i < m else 0
best = max(best, left + right + seq[n] * seq[i])
cache[key] = best
return best
def solve_11(seq, n, m, cache):
key = ('11', n, m)
if key in cache:
return cache[key]
assert m >= n + 2
if m == n + 2:
return max(seq[n] * seq[n+1] + seq[n+2], seq[n] + seq[n+1] * seq[n+2])
best = -1e9
for i in range(n + 1, m):
left = solve_11(seq, n, i, cache) if i > n + 1 else 0
right = solve_11(seq, i, m, cache) if i < m - 1 else 0
best = max(best, left + right + seq[i] * seq[n] + seq[m], left + right + seq[i] * seq[m] + seq[n])
cache[key] = best
return best
for c in [[1, 1, 1], [4, 2, 3, 5], [1, 2], [1, 2, 3], [1, 2, 3, 4, 5, 6, 7, 8, 9, 10]]:
print(c, solve_00(c, 0, len(c)-1, dict()))
Related
There is already a topic about this task, but I'd like to ask about my specific approach.
The task is:
Let A be a non-empty array consisting of N integers.
The abs sum of two for a pair of indices (P, Q) is the absolute value
|A[P] + A[Q]|, for 0 ≤ P ≤ Q < N.
For example, the following array A:
A[0] = 1 A1 = 4 A[2] = -3 has pairs of indices (0, 0), (0,
1), (0, 2), (1, 1), (1, 2), (2, 2). The abs sum of two for the pair
(0, 0) is A[0] + A[0] = |1 + 1| = 2. The abs sum of two for the pair
(0, 1) is A[0] + A1 = |1 + 4| = 5. The abs sum of two for the pair
(0, 2) is A[0] + A[2] = |1 + (−3)| = 2. The abs sum of two for the
pair (1, 1) is A1 + A1 = |4 + 4| = 8. The abs sum of two for the
pair (1, 2) is A1 + A[2] = |4 + (−3)| = 1. The abs sum of two for
the pair (2, 2) is A[2] + A[2] = |(−3) + (−3)| = 6. Write a function:
def solution(A)
that, given a non-empty array A consisting of N integers, returns the
minimal abs sum of two for any pair of indices in this array.
For example, given the following array A:
A[0] = 1 A1 = 4 A[2] = -3 the function should return 1, as
explained above.
Given array A:
A[0] = -8 A1 = 4 A[2] = 5 A[3] =-10 A[4] = 3 the
function should return |(−8) + 5| = 3.
Write an efficient algorithm for the following assumptions:
N is an integer within the range [1..100,000]; each element of array A
is an integer within the range [−1,000,000,000..1,000,000,000].
The official solution is O(N*M^2), but I think it could be solved in O(N).
My approach is to first get rid of duplicates and sort the array. Then we check both ends and sompare the abs sum moving the ends by one towards each other. We try to move the left end, the right one or both. If this doesn't improve the result, our sum is the lowest. My code is:
def solution(A):
A = list(set(A))
n = len(A)
A.sort()
beg = 0
end = n - 1
min_sum = abs(A[beg] + A[end])
while True:
min_left = abs(A[beg+1] + A[end]) if beg+1 < n else float('inf')
min_right = abs(A[beg] + A[end-1]) if end-1 >= 0 else float('inf')
min_both = abs(A[beg+1] + A[end-1]) if beg+1 < n and end-1 >= 0 else float('inf')
min_all = min([min_left, min_right, min_both])
if min_sum <= min_all:
return min_sum
if min_left == min_all:
beg += 1
min_sum = min_left
elif min_right == min_all:
end -= 1
min_sum = min_right
else:
beg += 1
end -= 1
min_sum = min_both
It passes almost all of the tests, but not all. Is there some bug in my code or the approach is wrong?
EDIT:
After the aka.nice answer I was able to fix the code. It scores 100% now.
def solution(A):
A = list(set(A))
n = len(A)
A.sort()
beg = 0
end = n - 1
min_sum = abs(A[beg] + A[end])
while beg <= end:
min_left = abs(A[beg+1] + A[end]) if beg+1 < n else float('inf')
min_right = abs(A[beg] + A[end-1]) if end-1 >= 0 else float('inf')
min_all = min(min_left, min_right)
if min_all < min_sum:
min_sum = min_all
if min_left <= min_all:
beg += 1
else:
end -= 1
return min_sum
Just take this example for array A
-11 -5 -2 5 6 8 12
and execute your algorithm step by step, you get a premature return:
beg=0
end=6
min_sum=1
min_left=7
min_right=3
min_both=3
min_all=3
return min_sum
though there is a better solution abs(5-5)=0.
Hint: you should check the sign of A[beg] and A[end] to decide whether to continue or exit the loop. What to do if both >= 0, if both <= 0, else ?
Note that A.sort() has a non neglectable cost, likely O(N*log(N)), it will dominate the cost of the solution you exhibit.
By the way, what is M in the official cost O(N*M^2)?
And the link you provide is another problem (sum all the elements of A or their opposite).
We are given an array with n values.
Example: [1,4,5,6,6]
For each index i of the array a ,we construct a new element of array b such that,
b[i]= [a[i]/1] + [a[i+1]/2] + [a[i+2]/3] + ⋯ + [a[n]/(n−i+1)] where [.] denotes the greatest integer function.
We are given an integer k as well.
We have to find the minimum i such that b[i] ≤ k.
I know the brute-force O(n^2) algorithm (to create the array - 'b'), can anybody suggest a better time complexity and way solve it?
For example, for the input [1,2,3],k=3, the output is 1(minimum-index).
Here, a[1]=1; a[2]=2; a[3]=3;
Now, b[1] = [a[1]/1] + [a[2]/2] + [a[3]/3] = [1/1] + [2/2] + [3/3] = 3;
b[2] = [a[2]/1] + [a[3]/2] = [2/1] + [3/2] = 3;
b[3] = [a[3]/1] = [3/1] = 3 (obvious)
Now, we have to find the index i such that b[i]<=k , k='3' , also b[1]<=3, henceforth, 1 is our answer! :-)
Constraints : - Time limits: -(2-seconds) , 1 <= a[i] <= 10^5, 1 <=
n <= 10^5, 1 <= k <= 10^9
Here's an O(n √A)-time algorithm to compute the b array where n is the number of elements in the a array and A is the maximum element of the a array.
This algorithm computes the difference sequence of the b array (∆b = b[0], b[1] - b[0], b[2] - b[1], ..., b[n-1] - b[n-2]) and derives b itself as the cumulative sums. Since the differences are linear, we can start with ∆b = 0, 0, ..., 0, loop over each element a[i], and add the difference sequence for [a[i]], [a[i]/2], [a[i]/3], ... at the appropriate spot. The key is that this difference sequence is sparse (less than 2√a[i] elements). For example, for a[i] = 36,
>>> [36//j for j in range(1,37)]
[36, 18, 12, 9, 7, 6, 5, 4, 4, 3, 3, 3, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]
>>> list(map(operator.sub,_,[0]+_[:-1]))
[36, -18, -6, -3, -2, -1, -1, -1, 0, -1, 0, 0, -1, 0, 0, 0, 0, 0, -1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]
We can derive the difference sequence from a subroutine that, given a positive integer r, returns all maximal pairs of positive integers (p, q) such that pq ≤ r.
See complete Python code below.
def maximal_pairs(r):
p = 1
q = r
while p < q:
yield (p, q)
p += 1
q = r // p
while q > 0:
p = r // q
yield (p, q)
q -= 1
def compute_b_fast(a):
n = len(a)
delta_b = [0] * n
for i, ai in enumerate(a):
previous_j = i
for p, q in maximal_pairs(ai):
delta_b[previous_j] += q
j = i + p
if j >= n:
break
delta_b[j] -= q
previous_j = j
for i in range(1, n):
delta_b[i] += delta_b[i - 1]
return delta_b
def compute_b_slow(a):
n = len(a)
b = [0] * n
for i, ai in enumerate(a):
for j in range(n - i):
b[i + j] += ai // (j + 1)
return b
for n in range(1, 100):
print(list(maximal_pairs(n)))
lst = [1, 34, 3, 2, 9, 21, 3, 2, 2, 1]
print(compute_b_fast(lst))
print(compute_b_slow(lst))
This probably cannot reach the efficiency of David Eisenstat's answer but since I spent quite a long time figuring out an implementation, I thought I'd leave it up anyway. As it is, it seems about O(n^2).
The elements of b[i] may be out of order, but sections of them are not:
[a[1]/1] + [a[2]/2] + [a[3]/3]
|------ s2_1 -----|
|-s1_1-|
[a[2]/1] + [a[3]/2]
|------ s2_2 -----|
|-s1_2-|
[a[3]/1]
|-s1_3-|
s2_1 < s2_2
s1_1 < s1_2 < s1_3
Binary search for k on s1. Any result with an s1_i greater than k will rule out a section of ordered rows (rows are b_is).
Binary search for k on s2 on the remaining rows. Any result with an s2_i greater than k will rule out a section of ordered rows (rows are b_is).
This wouldn't help much since in the worst case, we'd have O(n^2 * log n) complexity, greater than O(n^2).
But we can also search horizontally. If we know that b_i ≤ k, then it will rule out both all rows with greater or equal length and the need to search smaller s(m)s, not because smaller s(m)s cannot produce a sum >= k, but because they will necessarily produce one with a higher i and we are looking for the minimum i.
JavaScript code:
var sum_width_iterations = 0
var total_width_summed = 0
var sum_width_cache = {}
function sum_width(A, i, width){
let key = `${i},${width}`
if (sum_width_cache.hasOwnProperty(key))
return sum_width_cache[key]
sum_width_iterations++
total_width_summed += width
let result = 0
for (let j=A.length-width; j<A.length; j++)
result += ~~(A[j] / (j + 1 - i))
return sum_width_cache[key] = result
}
function get_b(A){
let result = []
A.map(function(a, i){
result.push(sum_width(A, i, A.length - i))
})
return result
}
function find_s_greater_than_k(A, width, low, high, k){
let mid = low + ((high - low) >> 1)
let s = sum_width(A, mid, width)
while (low <= high){
mid = low + ((high - low) >> 1)
s = sum_width(A, mid, width)
if (s > k)
high = mid - 1
else
low = mid + 1
}
return [mid, s]
}
function f(A, k, l, r){
let n = A.length
if (l > r){
console.log(`l > r: l, r: ${l}, ${r}`)
return [n + 1, Infinity]
}
let width = n - l
console.log(`\n(call) width, l, r: ${width}, ${l}, ${r}`)
let mid = l + ((r - l) >> 1)
let mid_width = n - mid
console.log(`mid: ${mid}`)
console.log('mid_width: ' + mid_width)
let highest_i = n - mid_width
let [i, s] = find_s_greater_than_k(A, mid_width, 0, highest_i, k)
console.log(`hi_i, s,i,k: ${highest_i}, ${s}, ${i}, ${k}`)
if (mid_width == width)
return [i, s]
// either way we need to look left
// and down
console.log(`calling left`)
let [li, ls] = f(A, k, l, mid - 1)
// if i is the highest, width is
// the width of b_i
console.log(`got left: li, ls, i, high_i: ${li}, ${ls}, ${i}, ${highest_i}`)
if (i == highest_i){
console.log(`i == highest_i, s <= k: ${s <= k}`)
// b_i is small enough
if (s <= k){
if (ls <= k)
return [li, ls]
else
return [i, s]
// b_i is larger than k
} else {
console.log(`b_i > k`)
let [ri, rs] = f(A, k, mid + 1, r)
console.log(`ri, rs: ${ri}, ${rs}`)
if (ls <= k)
return [li, ls]
else if (rs <= k)
return [ri, rs]
else
return [i, s]
}
// i < highest_i
} else {
console.log(`i < highest_i: high_i, i, s, li, ls, mid, mid_width, width, l, r: ${highest_i}, ${i}, ${s}, ${li}, ${ls}, ${mid}, ${mid_width}, ${width}, ${l}, ${r}`)
// get the full sum for this b
let b_i = sum_width(A, i, n - i)
console.log(`b_i: ${b_i}`)
// suffix sum is less than k
// so we cannot rule out either side
if (s < k){
console.log(`s < k`)
let ll = l
let lr = mid - 1
let [lli, lls] = f(A, k, ll, lr)
console.log(`ll, lr, lli, lls: ${ll}, ${lr}, ${lli}, ${lls}`)
// b_i is a match so we don't
// need to look to the right
if (b_i <= k){
console.log(`b_i <= k: i, b_i: ${i}, ${b_i}`)
if (lls <= k)
return [lli, lls]
else
return [i, b_i]
// b_i > k
} else {
console.log(`b_i > k: i, b_i: ${i}, ${b_i}`)
let rl = mid + 1
let rr = r
let [rri, rrs] = f(A, k, rl, rr)
console.log(`rl, rr, rri, rrs: ${rl}, ${rr}, ${rri}, ${rrs}`)
// return the best of right
// and left sections
if (lls <= k)
return [lli, lls]
else if (rrs <= k)
return [rri, rrs]
else
return [i, b_i]
}
// suffix sum is greater than or
// equal to k so we can rule out
// this and all higher rows (`b`s)
// that share this suffix
} else {
console.log(`s >= k`)
let ll = l
// the suffix rules out b_i
// and above
let lr = i - 1
let [lli, lls] = f(A, k, ll, lr)
console.log(`ll, lr, lli, lls: ${ll}, ${lr}, ${lli}, ${lls}`)
let rl = highest_i + 1
let rr = r
let [rri, rrs] = f(A, k, rl, rr)
console.log(`rl, rr, rri, rrs: ${rl}, ${rr}, ${rri}, ${rrs}`)
// return the best of right
// and left sections
if (lls <= k)
return [lli, lls]
else if (rrs <= k)
return [rri, rrs]
else
return [i, b_i]
}
}
}
let lst = [1, 2, 3, 1]
// b [3, 3, 3, 1]
lst = [ 1, 34, 3, 2, 9, 21, 3, 2, 2, 1]
// b [23, 41, 12, 13, 20, 22, 4, 3, 2, 1]
console.log(
JSON.stringify(f(lst, 20, 0, lst.length)))
console.log(`sum_width_iterations: ${sum_width_iterations}`)
console.log(`total_width_summed: ${total_width_summed}`)
Why should calculating b[i] lead to O(n²)? If i = 1, it takes n steps. If i = n, it takes one step to calculate b[i]...
You could improve your calculation when you abort the sum on the condition Sum > k.
Let a in N^n
Let k in N
for (i1 := 1; i1 <= n; i1++)
b := 0
for (i2 :=i1; i2 <= n; i2++) // This loop is the calculation of b[i]
b := b + ceil(a[i2]/(i2 + 1))
if (b > k)
break
if (i2 == n)
return i1
I am trying to understand the algorithm that solves this problem in O(log(n+m)) where n and m are the lengths of the arrays. I have taken the liberty to post the link to the explanation of this algorithm:
https://www.geeksforgeeks.org/median-of-two-sorted-arrays-of-different-sizes/
It's so hard for me to digest completely the idea behind this algorithm. I can see that the idea is to reduce the length of one of the arrays to either 1 or 2 and then apply the base cases. The base cases make sense, but I wonder if one can omit the base case for n = 2 and just work on n = 1. I also don't understand the remaining cases part. It looks so weird to me that we have to cut the array B[] from the start to idx. It's weird because idx can be equal to the length of B[], so we would be ignoring the whole array.
TL;DR:
The main idea is that you may delete N elements that are surely smaller than (or equal to) the median from your number set, as long as you delete the same amount that are surely greater or equal.
Let's explain it with an example:
A=[1 2 3 9 10], B=[3 4 5 6 7 8 9]
The middle elements marked:
A=[1 2 3 9 10], B=[3 4 5 6 7 8 9]
The overall median will be between 3 and 6, inclusive. So, if we delete two elements smaller than 3, and two elements greater than 6, we'll still have the same median. The smaller elements we delete from A, and the greater ones from B:
A=[3 9 10], B=[3 4 5 6 7]
Now we delete one element greater than 9 (from A) and one smaller than 5 (from B):
A=[3 9], B=[4 5 6 7]
We reached Case 4 (smaller array has 2 elements): the algorithm calls for the median of
B[M/2], B[M/2 – 1], max(A[0], B[M/2 – 2]), min(A[1], B[M/2 + 1])
being B[2], B[1], max(A[0], B[0]), min(A[1], B[3])
being 6, 5, max(3,4), min(9,7)
being [6 5 4 7]
The median of that array is 5.5. And that's the correct result.
def findmedian(A,B):
if len(A) > len(B):
return findmedian(B,A)# always ensuring that we do the binsearch on the shorter arr
x = len(A)
y = len(B)
start = 0# start and end of shorter arr
end = x
while (start <= end):
partition_x = (start + end)//2# the mid of smaller arr, partition_x is an index
partition_y = (x+y+1)//2 - partition_x# the suitable partition of larger arr to divide the arrs into equal halves
if partition_x == 0:# if there is nothing on the left
left_x = None
if partition_x == x:# if there is nothing on the right
right_x = sys.maxint# +inf
if partition_y == 0:
left_y = None# this is -inf similar to the case for smaller arr
if partition_y == y:
right_y = sys.maxint
if (left_x <= right_y) and (left_y <= right_x):# all the elems on left are smaller than all the elems on right is ensured by
#checking on the extremes only since arrs sorted. Also, the partition always makes equal halves, so found the right spot.
if (x+y) % 2 == 0:
return (max(left_x,left_y) + min(right_x,right_y))/2.0
else:
return max(left_x,left_y)# if the num of elems is odd
elif left_x > right_y:# if we have come more towards right of smaller arr, then move left on smaller arr
end = partition_x -1
else:# if we have come more to the left
start = partition_x + 1
class Solution(object):
def findMedianSortedArrays(self, nums1, nums2):
merged_array = (nums1 + nums2)
merged_array.sort()
l_m_a = len(merged_array)
count = int(l_m_a / 2)
if l_m_a % 2 == 1:
median = merged_array[count]
return median
else:
median_in_even = (merged_array[count] + merged_array[count - 1]) / 2
return median_in_even
class Solution:
def findMedianSortedArrays(self, nums1, nums2):
nums1.extend(nums2)
newArray = sorted(nums1)
if len(newArray)%2==0:
index = len(newArray)//2
median = (newArray[index] + newArray[index-1])/2
return float(median)
else:
index = len(newArray)//2
median = newArray[index]
return float(median)
if __name__ == '__main__':
obj = Solution()
print(obj.findMedianSortedArrays([1,3],[2]))
Median of two sorted Arrays | Same length | Different length
1st we need to merge both arrays in sorted order. And then we can find
the median. The Median will be the central element of the sorted array.
var findMedianSortedArrays = function(nums1, nums2) {
let array = [], leftIndex = 0, rightIndex = 0;
while (leftIndex < nums1.length && rightIndex < nums2.length) {
if (nums1[leftIndex] < nums2[rightIndex]) {
array.push(nums1[leftIndex]);
leftIndex++;
} else {
array.push(nums2[rightIndex]);
rightIndex++;
}
}
// add uninitiated remaining element from either array if any remains.
array = array.concat(nums1.slice(leftIndex)).concat(nums2.slice(rightIndex));
if (array.length % 2 == 0) {
return (array[(array.length / 2) - 1] + array[array.length / 2]) / 2;
} else {
return array[Math.floor(array.length / 2)]
}
};
findMedianSortedArrays([1 2 3 9 10], [3 4 5 6 7 8 9]);
/**
* #param {number[]} nums1
* #param {number[]} nums2
* #return {number}
*/
var findMedianSortedArrays = function (nums1, nums2) {
let newArray = [];
let median;
if (nums1.length > 0 && nums2.length > 0) {
newArray = [...nums1, ...nums2]
newArray.sort(function (a, b) {
return a - b;
})
} else if (nums1.length === 0) {
newArray = nums2
newArray.sort(function (a, b) {
return a - b;
})
} else if (nums2.length === 0) {
newArray = nums1
newArray.sort(function (a, b) {
return a - b;
})
}
if (newArray.length === 1) {
return newArray[0]
}
if (newArray.length === 3) {
return newArray[1]
}
if (newArray.length % 2 === 0) {
const findIndex = Math.floor(newArray.length / 2)
console.log("findIndexeven", findIndex)
const addValue = Math.max((newArray[findIndex - 1] +
newArray[findIndex]), 0)
median = addValue / 2
} else {
const findIndex = Math.floor(newArray.length / 2) + 1
console.log("findIndexodd", findIndex)
median = newArray[findIndex - 1]
}
console.log("medianValue",median)
return median
};
findMedianSortedArrays([1, 2], [3, 4])
For me, it's just several minutes of several lines of python codes and it passed the leetcode check with a runtime beating 62% of Python3 online submissions. My code is here:
class Solution:
def findMedianSortedArrays(self, nums1: List[int], nums2: List[int]) -> float:
n = len(nums1) + len(nums2)
nums1.extend(nums2) # merge two lists
nums1.sort() # sort it
if n % 2 == 0:
return (nums1[n//2-1] + nums1[n//2])/2 # return the median for even n
else:
return nums1[n//2] # return the median for odd n
I got this question as prerequisite for an interview,
A non-empty zero-indexed array A consisting of N integers is given. A
pit in this array is any triplet of integers (P, Q, R) such that: 0 ≤
P < Q < R < N;
sequence [A[P], A[P+1], ..., A[Q]] is strictly decreasing, i.e. A[P] >
A[P+1] > ... > A[Q];
sequence A[Q], A[Q+1], ..., A[R] is strictly increasing, i.e. A[Q] <
A[Q+1] < ... < A[R].
The depth of a pit (P, Q, R) is the number min{A[P] − A[Q], A[R] −
A[Q]}. For example, consider array A consisting of 10 elements such
that:
A[0] = 0
A[1] = 1
A[2] = 3
A[3] = -2
A[4] = 0
A[5] = 1
A[6] = 0
A[7] = -3
A[8] = 2
A[9] = 3
Triplet (2, 3, 4) is one of pits in this array, because sequence
[A[2], A[3]] is strictly decreasing (3 > −2) and sequence [A[3], A[4]]
is strictly increasing (−2 < 0). Its depth is min{A[2] − A[3], A[4] −
A[3]} = 2.
Triplet (2, 3, 5) is another pit with depth 3.
Triplet (5, 7, 8) is yet another pit with depth 4. There is no pit in
this array deeper (i.e. having depth greater) than 4.
It says that Triplet (5, 7, 8) has the deepest pit depth of 4.
but isn't Triplet (2, 7, 9) has the deepest pit depth 6?
corresponding value of Triplet (2, 7, 9) is (3, -3, 3) and it also satisfies the conditions mentioned, i.e.
1) 0 ≤ P < Q < R < N
2) A[P] > A[P+1] > ... > A[Q] and A[Q] < A[Q+1] < ... < A[R]
so in this case min{A[P] − A[Q], A[R] − A[Q]} is 6.
What am i missing here?
P.S. if you think this post does not belong here in this forum then please point out where should i post it.
See the sequence from P to Q for 2 to 7.
It is 3 -2 0 1 0 -3.
sequence [A[P], A[P+1], ..., A[Q]] is strictly decreasing, i.e. A[P] > A[P+1] > ... > A[Q];
The rule says that this should be a decreasing sequence. But it isn't. 3>-2 but -2 is not greater than 0. Here the sequence breaks.
From 7 to 9. No problem as the sequence is increasing. -3<2<3.
answer of the deepest pit problem in swift :
func solution(_ array: [Int]) -> Int {
//guaranty we have at least three elements
if array.isEmpty {
print("isEmpty")
return -1
}
if array.count < 3 {
print("is less than 3")
return -1
}
//extremum point; max or min points
var extremumPoints = [Int]()
//adding first element
extremumPoints.append(array[0])
//calculate extremum points for 1 to one before last element
for i in 1..<(array.count - 1) {
let isRelativeExtremum = ((array[i] - array[i - 1]) * (array[i] - array[i + 1])) > 0
//we call a point semi-extremum if a point is equal to previous element or next element and not equal to previous element or next element
let isSemiExtremum = ((array[i] != array[i - 1]) && (array[i] == array[i + 1])) || ((array[i] != array[i + 1]) && (array[i] == array[i - 1]))
if isRelativeExtremum || isSemiExtremum {
extremumPoints.append(array[i])
}
}
//adding last element
extremumPoints.append(array[array.count - 1])
//we will hold depthes in this array
var depthes = [Int]()
for i in 1..<(extremumPoints.count - 1) {
let isBottomOfaPit = extremumPoints[i] < extremumPoints[i - 1] && extremumPoints[i] < extremumPoints[i + 1]
if isBottomOfaPit {
let d1 = extremumPoints[i - 1] - extremumPoints[i]
let d2 = extremumPoints[i + 1] - extremumPoints[i]
let d = min(d1, d2)
depthes.append(d)
}
}
//deepest pit
let deepestPit = depthes.max()
return deepestPit ?? -1
}
//****************************
let A = [0,1,3,-2,0,1,0,-3,2,3]
let deepestPit = solution(A)
print(deepestPit) // 4
def deepest(A):
def check(p, q, r, A):
if A[p] > A[q] and A[q] < A[r]:
return min(A[p] - A[q], A[r] - A[q])
else:
return -1
max_depth = 0
for i in range(1, len(A) - 2):
if A[i-1] > A[i] < A[i + 1]:
p = i
r = i
while 0 <= p and r <= len(A) - 1:
depth = check(p, i, r, A)
max_depth = max(max_depth, depth)
p -= 1
r += 1
return max_depth
A question last week defined the zig zag ordering on an n by m matrix and asked how to list the elements in that order.
My question is how to quickly find the ith item in the zigzag ordering? That is, without traversing the matrix (for large n and m that's much too slow).
For example with n=m=8 as in the picture and (x, y) describing (row, column)
f(0) = (0, 0)
f(1) = (0, 1)
f(2) = (1, 0)
f(3) = (2, 0)
f(4) = (1, 1)
...
f(63) = (7, 7)
Specific question: what is the ten billionth (1e10) item in the zigzag ordering of a million by million matrix?
Let's assume that the desired element is located in the upper half of the matrix. The length of the diagonals are 1, 2, 3 ..., n.
Let's find the desired diagonal. It satisfies the following property:
sum(1, 2 ..., k) >= pos but sum(1, 2, ..., k - 1) < pos. The sum of 1, 2, ..., k is k * (k + 1) / 2. So we just need to find the smallest integer k such that k * (k + 1) / 2 >= pos. We can either use a binary search or solve this quadratic inequality explicitly.
When we know the k, we just need to find the pos - (k - 1) * k / 2 element of this diagonal. We know where it starts and where we should move(up or down, depending on the parity of k), so we can find the desired cell using a simple formula.
This solution has an O(1) or an O(log n) time complexity(it depends on whether we use a binary search or solve the inequation explicitly in step 2).
If the desired element is located in the lower half of the matrix, we can solve this problem for a pos' = n * n - pos + 1 and then use symmetry to get the solution to the original problem.
I used 1-based indexing in this solution, using 0-based indexing might require adding +1 or -1 somewhere, but the idea of the solution is the same.
If the matrix is rectangular, not square, we need to consider the fact the length of diagonals look this way: 1, 2, 3, ..., m, m, m, .., m, m - 1, ..., 1(if m <= n) when we search for the k, so the sum becomes something like k * (k + 1) / 2 if k <= m and k * (k + 1) / 2 + m * (k - m) otherwise.
import math, random
def naive(n, m, ord, swap = False):
dx = 1
dy = -1
if swap:
dx, dy = dy, dx
cur = [0, 0]
for i in range(ord):
cur[0] += dy
cur[1] += dx
if cur[0] < 0 or cur[1] < 0 or cur[0] >= n or cur[1] >= m:
dx, dy = dy, dx
if cur[0] >= n:
cur[0] = n - 1
cur[1] += 2
if cur[1] >= m:
cur[1] = m - 1
cur[0] += 2
if cur[0] < 0: cur[0] = 0
if cur[1] < 0: cur[1] = 0
return cur
def fast(n, m, ord, swap = False):
if n < m:
x, y = fast(m, n, ord, not swap)
return [y, x]
alt = n * m - ord - 1
if alt < ord:
x, y = fast(n, m, alt, swap if (n + m) % 2 == 0 else not swap)
return [n - x - 1, m - y - 1]
if ord < (m * (m + 1) / 2):
diag = int((-1 + math.sqrt(1 + 8 * ord)) / 2)
parity = (diag + (0 if swap else 1)) % 2
within = ord - (diag * (diag + 1) / 2)
if parity: return [diag - within, within]
else: return [within, diag - within]
else:
ord -= (m * (m + 1) / 2)
diag = int(ord / m)
within = ord - diag * m
diag += m
parity = (diag + (0 if swap else 1)) % 2
if not parity:
within = m - within - 1
return [diag - within, within]
if __name__ == "__main__":
for i in range(1000):
n = random.randint(3, 100)
m = random.randint(3, 100)
ord = random.randint(0, n * m - 1)
swap = random.randint(0, 99) < 50
na = naive(n, m, ord, swap)
fa = fast(n, m, ord, swap)
assert na == fa, "(%d, %d, %d, %s) ==> (%s), (%s)" % (n, m, ord, swap, na, fa)
print fast(1000000, 1000000, 9999999999, False)
print fast(1000000, 1000000, 10000000000, False)
So the 10-billionth element (the one with ordinal 9999999999), and the 10-billion-first element (the one with ordinal 10^10) are:
[20331, 121089]
[20330, 121090]
An analytical solution
In the general case, your matrix will be divided in 3 areas:
an initial triangle t1
a skewed part mid where diagonals have a constant length
a final triangle t2
Let's call p the index of your diagonal run.
We want to define two functions x(p) and y(p) that give you the column and row of the pth cell.
Initial triangle
Let's look at the initial triangular part t1, where each new diagonal is one unit longer than the preceding.
Now let's call d the index of the diagonal that holds the cell, and
Sp = sum(di) for i in [0..p-1]
We have p = Sp + k, with 0 <=k <= d and
Sp = d(d+1)/2
if we solve for d, it brings
d²+d-2p = 0, a quadratic equation where we retain only the positive root:
d = (-1+sqrt(1+8*p))/2
Now we want the highest integer value closest to d, which is floor(d).
In the end, we have
p = d + k with d = floor((-1+sqrt(1+8*p))/2) and k = p - d(d+1)/2
Let's call
o(d) the function that equals 1 if d is odd and 0 otherwise, and
e(d) the function that equals 1 if d is even and 0 otherwise.
We can compute x(p) and y(p) like so:
d = floor((-1+sqrt(1+8*p))/2)
k = p - d(d+1)/2
o = d % 2
e = 1 - o
x = e*d + (o-e)*k
y = o*d + (e-o)*k
even and odd functions are used to try to salvage some clarity, but you can replace
e(p) with 1 - o(p) and have slightly more efficient but less symetric formulaes for x and y.
Middle part
let's consider the smallest matrix dimension s, i.e. s = min (m,n).
The previous formulaes hold until x or y (whichever comes first) reaches the value s.
The upper bound of p such as x(i) <= s and y(i) <= s for all i in [0..p]
(i.e. the cell indexed by p is inside the initial triangle t1) is given by
pt1 = s(s+1)/2.
For p >= pt1, diagonal length remains equal to s until we reach the second triangle t2.
when inside mid, we have:
p = s(s+1)/2 + ds + k with k in [0..s[.
which yields:
d = floor ((p - s(s+1)/2)/s)
k = p - ds
We can then use the same even/odd trick to compute x(p) and y(p):
p -= s(s+1)/2
d = floor (p / s)
k = p - d*s
o = (d+s) % 2
e = 1 - o
x = o*s + (e-o)*k
y = e*s + (o-e)*k
if (n > m)
x += d+e
y -= e
else
y += d+o
x -= o
Final triangle
Using symetry, we can calculate pt2 = m*n - s(s+1)/2
We now face nearly the same problem as for t1, except that the diagonal may run in the same direction as for t1 or in the reverse direction (if n+m is odd).
Using symetry tricks, we can compute x(p) and y(p) like so:
p = n*m -1 - p
d = floor((-1+sqrt(1+8*p))/2)
k = p - d*(d+1)/2
o = (d+m+n) % 2
e = 1 - $o;
x = n-1 - (o*d + (e-o)*k)
y = m-1 - (e*d + (o-e)*k)
Putting all together
Here is a sample c++ implementation.
I used 64 bits integers out of sheer lazyness. Most could be replaced by 32 bits values.
The computations could be made more effective by precomputing a few more coefficients.
A good part of the code could be factorized, but I doubt it is worth the effort.
Since this is just a quick and dirty proof of concept, I did not optimize it.
#include <cstdio> // printf
#include <algorithm> // min
using namespace std;
typedef long long tCoord;
void panic(const char * msg)
{
printf("PANIC: %s\n", msg);
exit(-1);
}
struct tPoint {
tCoord x, y;
tPoint(tCoord x = 0, tCoord y = 0) : x(x), y(y) {}
tPoint operator+(const tPoint & p) const { return{ x + p.x, y + p.y }; }
bool operator!=(const tPoint & p) const { return x != p.x || y != p.y; }
};
class tMatrix {
tCoord n, m; // dimensions
tCoord s; // smallest dimension
tCoord pt1, pt2; // t1 / mid / t2 limits for p
public:
tMatrix(tCoord n, tCoord m) : n(n), m(m)
{
s = min(n, m);
pt1 = (s*(s + 1)) / 2;
pt2 = n*m - pt1;
}
tPoint diagonal_cell(tCoord p)
{
tCoord x, y;
if (p < pt1) // inside t1
{
tCoord d = (tCoord)floor((-1 + sqrt(1 + 8 * p)) / 2);
tCoord k = p - (d*(d + 1)) / 2;
tCoord o = d % 2;
tCoord e = 1 - o;
x = o*d + (e - o)*k;
y = e*d + (o - e)*k;
}
else if (p < pt2) // inside mid
{
p -= pt1;
tCoord d = (tCoord)floor(p / s);
tCoord k = p - d*s;
tCoord o = (d + s) % 2;
tCoord e = 1 - o;
x = o*s + (e - o)*k;
y = e*s + (o - e)*k;
if (m > n) // vertical matrix
{
x -= o;
y += d + o;
}
else // horizontal matrix
{
x += d + e;
y -= e;
}
}
else // inside t2
{
p = n * m - 1 - p;
tCoord d = (tCoord)floor((-1 + sqrt(1 + 8 * p)) / 2);
tCoord k = p - (d*(d + 1)) / 2;
tCoord o = (d + m + n) % 2;
tCoord e = 1 - o;
x = n - 1 - (o*d + (e - o)*k);
y = m - 1 - (e*d + (o - e)*k);
}
return{ x, y };
}
void check(void)
{
tPoint move[4] = { { 1, 0 }, { -1, 1 }, { 1, -1 }, { 0, 1 } };
tPoint pos;
tCoord dir = 0;
for (tCoord p = 0; p != n * m ; p++)
{
tPoint dc = diagonal_cell(p);
if (pos != dc) panic("zot!");
pos = pos + move[dir];
if (dir == 0)
{
if (pos.y == m - 1) dir = 2;
else dir = 1;
}
else if (dir == 3)
{
if (pos.x == n - 1) dir = 1;
else dir = 2;
}
else if (dir == 1)
{
if (pos.y == m - 1) dir = 0;
else if (pos.x == 0) dir = 3;
}
else
{
if (pos.x == n - 1) dir = 3;
else if (pos.y == 0) dir = 0;
}
}
}
};
void main(void)
{
const tPoint dim[] = { { 10, 10 }, { 11, 11 }, { 10, 30 }, { 30, 10 }, { 10, 31 }, { 31, 10 }, { 11, 31 }, { 31, 11 } };
for (tPoint d : dim)
{
printf("Checking a %lldx%lld matrix...", d.x, d.y);
tMatrix(d.x, d.y).check();
printf("done\n");
}
tCoord p = 10000000000;
tMatrix matrix(1000000, 1000000);
tPoint cell = matrix.diagonal_cell(p);
printf("Coordinates of %lldth cell: (%lld,%lld)\n", p, cell.x, cell.y);
}
Results are checked against "manual" sweep of the matrix.
This "manual" sweep is a ugly hack that won't work for a one-row or one-column matrix, though diagonal_cell() does work on any matrix (the "diagonal" sweep becomes linear in that case).
The coordinates found for the 10.000.000.000th cell of a 1.000.000x1.000.000 matrix seem consistent, since the diagonal d on which the cell stands is about sqrt(2*1e10), approx. 141421, and the sum of cell coordinates is about equal to d (121090+20330 = 141420). Besides, it is also what the two other posters report.
I would say there is a good chance this lump of obfuscated code actually produces an O(1) solution to your problem.