Recently I needed to do weighted random selection of elements from a list, both with and without replacement. While there are well known and good algorithms for unweighted selection, and some for weighted selection without replacement (such as modifications of the resevoir algorithm), I couldn't find any good algorithms for weighted selection with replacement. I also wanted to avoid the resevoir method, as I was selecting a significant fraction of the list, which is small enough to hold in memory.
Does anyone have any suggestions on the best approach in this situation? I have my own solutions, but I'm hoping to find something more efficient, simpler, or both.
One of the fastest ways to make many with replacement samples from an unchanging list is the alias method. The core intuition is that we can create a set of equal-sized bins for the weighted list that can be indexed very efficiently through bit operations, to avoid a binary search. It will turn out that, done correctly, we will need to only store two items from the original list per bin, and thus can represent the split with a single percentage.
Let's us take the example of five equally weighted choices, (a:1, b:1, c:1, d:1, e:1)
To create the alias lookup:
Normalize the weights such that they sum to 1.0. (a:0.2 b:0.2 c:0.2 d:0.2 e:0.2) This is the probability of choosing each weight.
Find the smallest power of 2 greater than or equal to the number of variables, and create this number of partitions, |p|. Each partition represents a probability mass of 1/|p|. In this case, we create 8 partitions, each able to contain 0.125.
Take the variable with the least remaining weight, and place as much of it's mass as possible in an empty partition. In this example, we see that a fills the first partition. (p1{a|null,1.0},p2,p3,p4,p5,p6,p7,p8) with (a:0.075, b:0.2 c:0.2 d:0.2 e:0.2)
If the partition is not filled, take the variable with the most weight, and fill the partition with that variable.
Repeat steps 3 and 4, until none of the weight from the original partition need be assigned to the list.
For example, if we run another iteration of 3 and 4, we see
(p1{a|null,1.0},p2{a|b,0.6},p3,p4,p5,p6,p7,p8) with (a:0, b:0.15 c:0.2 d:0.2 e:0.2) left to be assigned
At runtime:
Get a U(0,1) random number, say binary 0.001100000
bitshift it lg2(p), finding the index partition. Thus, we shift it by 3, yielding 001.1, or position 1, and thus partition 2.
If the partition is split, use the decimal portion of the shifted random number to decide the split. In this case, the value is 0.5, and 0.5 < 0.6, so return a.
Here is some code and another explanation, but unfortunately it doesn't use the bitshifting technique, nor have I actually verified it.
A simple approach that hasn't been mentioned here is one proposed in Efraimidis and Spirakis. In python you could select m items from n >= m weighted items with strictly positive weights stored in weights, returning the selected indices, with:
import heapq
import math
import random
def WeightedSelectionWithoutReplacement(weights, m):
elt = [(math.log(random.random()) / weights[i], i) for i in range(len(weights))]
return [x[1] for x in heapq.nlargest(m, elt)]
This is very similar in structure to the first approach proposed by Nick Johnson. Unfortunately, that approach is biased in selecting the elements (see the comments on the method). Efraimidis and Spirakis proved that their approach is equivalent to random sampling without replacement in the linked paper.
Here's what I came up with for weighted selection without replacement:
def WeightedSelectionWithoutReplacement(l, n):
"""Selects without replacement n random elements from a list of (weight, item) tuples."""
l = sorted((random.random() * x[0], x[1]) for x in l)
return l[-n:]
This is O(m log m) on the number of items in the list to be selected from. I'm fairly certain this will weight items correctly, though I haven't verified it in any formal sense.
Here's what I came up with for weighted selection with replacement:
def WeightedSelectionWithReplacement(l, n):
"""Selects with replacement n random elements from a list of (weight, item) tuples."""
cuml = []
total_weight = 0.0
for weight, item in l:
total_weight += weight
cuml.append((total_weight, item))
return [cuml[bisect.bisect(cuml, random.random()*total_weight)] for x in range(n)]
This is O(m + n log m), where m is the number of items in the input list, and n is the number of items to be selected.
I'd recommend you start by looking at section 3.4.2 of Donald Knuth's Seminumerical Algorithms.
If your arrays are large, there are more efficient algorithms in chapter 3 of Principles of Random Variate Generation by John Dagpunar. If your arrays are not terribly large or you're not concerned with squeezing out as much efficiency as possible, the simpler algorithms in Knuth are probably fine.
It is possible to do Weighted Random Selection with replacement in O(1) time, after first creating an additional O(N)-sized data structure in O(N) time. The algorithm is based on the Alias Method developed by Walker and Vose, which is well described here.
The essential idea is that each bin in a histogram would be chosen with probability 1/N by a uniform RNG. So we will walk through it, and for any underpopulated bin which would would receive excess hits, assign the excess to an overpopulated bin. For each bin, we store the percentage of hits which belong to it, and the partner bin for the excess. This version tracks small and large bins in place, removing the need for an additional stack. It uses the index of the partner (stored in bucket[1]) as an indicator that they have already been processed.
Here is a minimal python implementation, based on the C implementation here
def prep(weights):
data_sz = len(weights)
factor = data_sz/float(sum(weights))
data = [[w*factor, i] for i,w in enumerate(weights)]
big=0
while big<data_sz and data[big][0]<=1.0: big+=1
for small,bucket in enumerate(data):
if bucket[1] is not small: continue
excess = 1.0 - bucket[0]
while excess > 0:
if big==data_sz: break
bucket[1] = big
bucket = data[big]
bucket[0] -= excess
excess = 1.0 - bucket[0]
if (excess >= 0):
big+=1
while big<data_sz and data[big][0]<=1: big+=1
return data
def sample(data):
r=random.random()*len(data)
idx = int(r)
return data[idx][1] if r-idx > data[idx][0] else idx
Example usage:
TRIALS=1000
weights = [20,1.5,9.8,10,15,10,15.5,10,8,.2];
samples = [0]*len(weights)
data = prep(weights)
for _ in range(int(sum(weights)*TRIALS)):
samples[sample(data)]+=1
result = [float(s)/TRIALS for s in samples]
err = [a-b for a,b in zip(result,weights)]
print(result)
print([round(e,5) for e in err])
print(sum([e*e for e in err]))
The following is a description of random weighted selection of an element of a
set (or multiset, if repeats are allowed), both with and without replacement in O(n) space
and O(log n) time.
It consists of implementing a binary search tree, sorted by the elements to be
selected, where each node of the tree contains:
the element itself (element)
the un-normalized weight of the element (elementweight), and
the sum of all the un-normalized weights of the left-child node and all of
its children (leftbranchweight).
the sum of all the un-normalized weights of the right-child node and all of
its chilren (rightbranchweight).
Then we randomly select an element from the BST by descending down the tree. A
rough description of the algorithm follows. The algorithm is given a node of
the tree. Then the values of leftbranchweight, rightbranchweight,
and elementweight of node is summed, and the weights are divided by this
sum, resulting in the values leftbranchprobability,
rightbranchprobability, and elementprobability, respectively. Then a
random number between 0 and 1 (randomnumber) is obtained.
if the number is less than elementprobability,
remove the element from the BST as normal, updating leftbranchweight
and rightbranchweight of all the necessary nodes, and return the
element.
else if the number is less than (elementprobability + leftbranchweight)
recurse on leftchild (run the algorithm using leftchild as node)
else
recurse on rightchild
When we finally find, using these weights, which element is to be returned, we either simply return it (with replacement) or we remove it and update relevant weights in the tree (without replacement).
DISCLAIMER: The algorithm is rough, and a treatise on the proper implementation
of a BST is not attempted here; rather, it is hoped that this answer will help
those who really need fast weighted selection without replacement (like I do).
This is an old question for which numpy now offers an easy solution so I thought I would mention it. Current version of numpy is version 1.2 and numpy.random.choice allows the sampling to be done with or without replacement and with given weights.
Suppose you want to sample 3 elements without replacement from the list ['white','blue','black','yellow','green'] with a prob. distribution [0.1, 0.2, 0.4, 0.1, 0.2]. Using numpy.random module it is as easy as this:
import numpy.random as rnd
sampling_size = 3
domain = ['white','blue','black','yellow','green']
probs = [.1, .2, .4, .1, .2]
sample = rnd.choice(domain, size=sampling_size, replace=False, p=probs)
# in short: rnd.choice(domain, sampling_size, False, probs)
print(sample)
# Possible output: ['white' 'black' 'blue']
Setting the replace flag to True, you have a sampling with replacement.
More info here:
http://docs.scipy.org/doc/numpy/reference/generated/numpy.random.choice.html#numpy.random.choice
We faced a problem to randomly select K validators of N candidates once per epoch proportionally to their stakes. But this gives us the following problem:
Imagine probabilities of each candidate:
0.1
0.1
0.8
Probabilities of each candidate after 1'000'000 selections 2 of 3 without replacement became:
0.254315
0.256755
0.488930
You should know, those original probabilities are not achievable for 2 of 3 selection without replacement.
But we wish initial probabilities to be a profit distribution probabilities. Else it makes small candidate pools more profitable. So we realized that random selection with replacement would help us – to randomly select >K of N and store also weight of each validator for reward distribution:
std::vector<int> validators;
std::vector<int> weights(n);
int totalWeights = 0;
for (int j = 0; validators.size() < m; j++) {
int value = rand() % likehoodsSum;
for (int i = 0; i < n; i++) {
if (value < likehoods[i]) {
if (weights[i] == 0) {
validators.push_back(i);
}
weights[i]++;
totalWeights++;
break;
}
value -= likehoods[i];
}
}
It gives an almost original distribution of rewards on millions of samples:
0.101230
0.099113
0.799657
Related
I'm working with arrays of integer, all of the same size l.
I have a static set of them and I need to build a function to efficiently look them up.
The tricky part is that the elements in the array I need to search might be off by 1.
Given the arrays {A_1, A_2, ..., A_n}, and an array S, I need a function search such that:
search(S)=x iff ∀i: A_x[i] ∈ {S[i]-1, S[i], S[i]+1}.
A possible solution is treating each vector as a point in an l-dimensional space and looking for the closest point, but it'd cost something like O(l*n) in space and O(l*log(n)) in time.
Would there be a solution with a better space complexity (and/or time, of course)?
My arrays are pretty different from each other, and good heuristics might be enough.
Consider a search array S with the values:
S = [s1, s2, s3, ... , sl]
and the average value:
s̅ = (s1 + s2 + s3 + ... + sl) / l
and two matching arrays, one where every value is one greater than the corresponding value in S, and one where very value is one smaller:
A1 = [s1+1, s2+1, s3+1, ... , sl+1]
A2 = [s1−1, s2−1, s3−1, ... , sl−1]
These two arrays would have the average values:
a̅1 = (s1 + 1 + s2 + 1 + s3 + 1 + ... + sl + 1) / l = s̅ + 1
a̅2 = (s1 − 1 + s2 − 1 + s3 − 1 + ... + sl − 1) / l = s̅ − 1
So every matching array, whose values are at most 1 away from the corresponding values in the search array, has an average value that is at most 1 away from the average value of the search array.
If you calculate and store the average value of each array, and then sort the arrays based on their average value (or use an extra data structure that enables you to find all arrays with a certain average value), you can quickly identify which arrays have an average value within 1 of the search array's average value. Depending on the data, this could drastically reduce the number of arrays you have to check for similarity.
After having pre-processed the arrays and stores their average values, performing a search would mean iterating over the search array to calculate the average value, looking up which arrays have a similar average value, and then iterating over those arrays to check every value.
If you expect many arrays to have a similar average value, you could use several averages to detect arrays that are locally very different but similar on average. You could e.g. calculate these four averages:
the first half of the array
the second half of the array
the odd-numbered elements
the even-numbered elements
Analysis of the actual data should give you more information about how to divide the array and combine different averages to be most effective.
If the total sum of an array cannot exceed the integer size, you could store the total sum of each array, and check whether it is within l of the total sum of the search array, instead of using averages. This would avoid having to use floats and divisions.
(You could expand this idea by also storing other properties which are easily calculated and don't take up much space to store, such as the highest and lowest value, the biggest jump, ... They could help create a fingerprint of each array that is near-unique, depending on the data.)
If the number of dimensions is not very small, then probably the best solution will be to build a decision tree that recursively partitions the set along different dimensions.
Each node, including the root, would be a hash table from the possible values for some dimension to either:
The list of points that match that value within tolerance, if it's small enough; or
Those same points in a similar tree partitioning on the remaining dimensions.
Since each level completely eliminates one dimension, the depth of the tree is at most L, and search takes O(L) time.
The order in which the dimensions are chosen along each path is important, of course -- the wrong choice could explode the size of the data structure, with each point appearing many times.
Since your points are "pretty different", though, it should be possible to build a tree with minimal duplication. I would try the ID3 algorithm to choose the dimensions: https://en.wikipedia.org/wiki/ID3_algorithm. That basically means you greedily choose the dimension that maximizes the overall reduction in set size, using an entropy metric.
I would personally create something like a Trie for the lookup. I said "something like" because we have up to 3 values per index that might match. So we aren't creating a decision tree, but a DAG. Where sometimes we have choices.
That is straightforward and will run (with backtracking) in maximum time O(k*l).
But here is the trick. Whenever we see a choice of matching states that we can go into next, we can create a merged state which tries all of them. We can create a few or a lot of these merged states. Each one will defer a choice by 1 step. And if we're careful to keep track of which merged states we've created, we can reuse the same one over and over again.
In theory we can be generating partial matches for somewhat arbitrary subsets of our arrays. Which can grow exponentially in the number of arrays. In practice are likely to only wind up with a few of these merged states. But still we can guarantee a tradeoff - more states up front runs faster later. So we optimize until we are done or have hit the limit of how much data we want to have.
Here is some proof of concept code for this in Python. It will likely build the matcher in time O(n*l) and match in time O(l). However it is only guaranteed to build the matcher in time O(n^2 * l^2) and match in time O(n * l).
import pprint
class Matcher:
def __init__ (self, arrays, optimize_limit=None):
# These are the partial states we could be in during a match.
self.states = [{}]
# By state, this is what we would be trying to match.
self.state_for = ['start']
# By combination we could try to match for, which state it is.
self.comb_state = {'start': 0}
for i in range(len(arrays)):
arr = arrays[i]
# Set up "matched the end".
state_index = len(self.states)
this_state = {'matched': [i]}
self.comb_state[(i, len(arr))] = state_index
self.states.append(this_state)
self.state_for.append((i, len(arr)))
for j in reversed(range(len(arr))):
this_for = (i, j)
prev_state = {}
if 0 == j:
prev_state = self.states[0]
matching_values = set((arr[k] for k in range(max(j-1, 0), min(j+2, len(arr)))))
for v in matching_values:
if v in prev_state:
prev_state[v].append(state_index)
else:
prev_state[v] = [state_index]
if 0 < j:
state_index = len(self.states)
self.states.append(prev_state)
self.state_for.append(this_for)
self.comb_state[this_for] = state_index
# Theoretically optimization can take space
# O(2**len(arrays) * len(arrays[0]))
# We will optimize until we are done or hit a more reasonable limit.
if optimize_limit is None:
# Normally
optimize_limit = len(self.states)**2
# First we find all of the choices at the root.
# This will be an array of arrays with format:
# [state, key, values]
todo = []
for k, v in self.states[0].iteritems():
if 1 < len(v):
todo.append([self.states[0], k, tuple(v)])
while len(todo) and len(self.states) < optimize_limit:
this_state, this_key, this_match = todo.pop(0)
if this_key == 'matched':
pass # We do not need to optimize this!
elif this_match in self.comb_state:
this_state[this_key] = self.comb_state[this_match]
else:
# Construct a new state that is all of these.
new_state = {}
for state_ind in this_match:
for k, v in self.states[state_ind].iteritems():
if k in new_state:
new_state[k] = new_state[k] + v
else:
new_state[k] = v
i = len(self.states)
self.states.append(new_state)
self.comb_state[this_match] = i
self.state_for.append(this_match)
this_state[this_key] = [i]
for k, v in new_state.iteritems():
if 1 < len(v):
todo.append([new_state, k, tuple(v)])
#pp = pprint.PrettyPrinter()
#pp.pprint(self.states)
#pp.pprint(self.comb_state)
#pp.pprint(self.state_for)
def match (self, list1, ind=0, state=0):
this_state = self.states[state]
if 'matched' in this_state:
return this_state['matched']
elif list1[ind] in this_state:
answer = []
for next_state in this_state[list1[ind]]:
answer = answer + self.match(list1, ind+1, next_state)
return answer;
else:
return []
foo = Matcher([[1, 2, 3], [2, 3, 4]])
print(foo.match([2, 2, 3]))
Please note that I deliberately set up a situation where there are 2 matches. It reports both of them. :-)
I came up with a further approach derived off Matt Timmermans's answer: building a simple decision tree that might have certain some arrays in multiple branches. It works even if the error in the array I'm searching is larger than 1.
The idea is the following: given the set of arrays As...
Pick an index and a pivot.
I fixed the pivot to a constant value that works well with my data, and tried all indices to find the best one. Trying multiple pivots might work better, but I didn't need to.
Partition As into two possibly-intersecting subsets, one for the arrays (whose index-th element is) smaller than the pivot, one for the larger arrays. Arrays very close to the pivot are added to both sets:
function partition( As, pivot, index ):
return {
As.filter( A => A[index] <= pivot + 1 ),
As.filter( A => A[index] >= pivot - 1 ),
}
Apply both previous steps to each subset recursively, stopping when a subset only contains a single element.
Here an example of a possible tree generated with this algorithm (note that A2 appears both on the left and right child of the root node):
{A1, A2, A3, A4}
pivot:15
index:73
/ \
/ \
{A1, A2} {A2, A3, A4}
pivot:7 pivot:33
index:54 index:0
/ \ / \
/ \ / \
A1 A2 {A2, A3} A4
pivot:5
index:48
/ \
/ \
A2 A3
The search function then uses this as a normal decision tree: it starts from the root node and recurses either to the left or the right child depending on whether its value at index currentNode.index is greater or less than currentNode.pivot. It proceeds recursively until it reaches a leaf.
Once the decision tree is built, the time complexity is in the worst case O(n), but in practice it's probably closer to O(log(n)) if we choose good indices and pivots (and if the dataset is diverse enough) and find a fairly balanced tree.
The space complexity can be really bad in the worst case (O(2^n)), but it's closer to O(n) with balanced trees.
I have a set of points (x,y).
i need to return two points with minimal distance.
I use this:
http://www.cs.ucsb.edu/~suri/cs235/ClosestPair.pdf
but , i dont really understand how the algo is working.
Can explain in more simple how the algo working?
or suggest another idea?
Thank!
If the number of points is small, you can use the brute force approach i.e:
for each point find the closest point among other points and save the minimum distance with the current two indices till now.
If the number of points is large, I think you may find the answer in this thread:
Shortest distance between points algorithm
Solution for Closest Pair Problem with minimum time complexity O(nlogn) is divide-and-conquer methodology as it mentioned in the document that you have read.
Divide-and-conquer Approach for Closest-Pair Problem
Easiest way to understand this algorithm is reading an implementation of it in a high-level language (because sometimes understanding the algorithms or pseudo-codes can be harder than understanding the real codes) like Python:
# closest pairs by divide and conquer
# David Eppstein, UC Irvine, 7 Mar 2002
from __future__ import generators
def closestpair(L):
def square(x): return x*x
def sqdist(p,q): return square(p[0]-q[0])+square(p[1]-q[1])
# Work around ridiculous Python inability to change variables in outer scopes
# by storing a list "best", where best[0] = smallest sqdist found so far and
# best[1] = pair of points giving that value of sqdist. Then best itself is never
# changed, but its elements best[0] and best[1] can be.
#
# We use the pair L[0],L[1] as our initial guess at a small distance.
best = [sqdist(L[0],L[1]), (L[0],L[1])]
# check whether pair (p,q) forms a closer pair than one seen already
def testpair(p,q):
d = sqdist(p,q)
if d < best[0]:
best[0] = d
best[1] = p,q
# merge two sorted lists by y-coordinate
def merge(A,B):
i = 0
j = 0
while i < len(A) or j < len(B):
if j >= len(B) or (i < len(A) and A[i][1] <= B[j][1]):
yield A[i]
i += 1
else:
yield B[j]
j += 1
# Find closest pair recursively; returns all points sorted by y coordinate
def recur(L):
if len(L) < 2:
return L
split = len(L)/2
L = list(merge(recur(L[:split]), recur(L[split:])))
# Find possible closest pair across split line
# Note: this is not quite the same as the algorithm described in class, because
# we use the global minimum distance found so far (best[0]), instead of
# the best distance found within the recursive calls made by this call to recur().
for i in range(len(E)):
for j in range(1,8):
if i+j < len(E):
testpair(E[i],E[i+j])
return L
L.sort()
recur(L)
return best[1]
closestpair([(0,0),(7,6),(2,20),(12,5),(16,16),(5,8),\
(19,7),(14,22),(8,19),(7,29),(10,11),(1,13)])
# returns: (7,6),(5,8)
Taken from: https://www.ics.uci.edu/~eppstein/161/python/closestpair.py
Detailed explanation:
First we define an Euclidean distance aka Square distance function to prevent code repetition.
def square(x): return x*x # Define square function
def sqdist(p,q): return square(p[0]-q[0])+square(p[1]-q[1]) # Define Euclidean distance function
Then we are taking the first two points as our initial best guess:
best = [sqdist(L[0],L[1]), (L[0],L[1])]
This is a function definition for comparing Euclidean distances of next pair with our current best pair:
def testpair(p,q):
d = sqdist(p,q)
if d < best[0]:
best[0] = d
best[1] = p,q
def merge(A,B): is just a rewind function for the algorithm to merge two sorted lists that previously divided to half.
def recur(L): function definition is the actual body of the algorithm. So I will explain this function definition in more detail:
if len(L) < 2:
return L
with this part, algorithm terminates the recursion if there is only one element/point left in the list of points.
Split the list to half: split = len(L)/2
Create a recursion (by calling function's itself) for each half: L = list(merge(recur(L[:split]), recur(L[split:])))
Then lastly this nested loops will test whole pairs in the current half-list with each other:
for i in range(len(E)):
for j in range(1,8):
if i+j < len(E):
testpair(E[i],E[i+j])
As the result of this, if a better pair is found best pair will be updated.
So they solve for the problem in Many dimensions using a divide-and-conquer approach. Binary search or divide-and-conquer is mega fast. Basically, if you can split a dataset into two halves, and keep doing that until you find some info you want, you are doing it as fast as humanly and computerly possible most of the time.
For this question, it means that we divide the data set of points into two sets, S1 and S2.
All the points are numerical, right? So we have to pick some number where to divide the dataset.
So we pick some number m and say it is the median.
So let's take a look at an example:
(14, 2)
(11, 2)
(5, 2)
(15, 2)
(0, 2)
What's the closest pair?
Well, they all have the same Y coordinate, so we can look at Xs only... X shortest distance is 14 to 15, a distance of 1.
How can we figure that out using divide-and-conquer?
We look at the greatest value of X and the smallest value of X and we choose the median as a dividing line to make our two sets.
Our median is 7.5 in this example.
We then make 2 sets
S1: (0, 2) and (5, 2)
S2: (11, 2) and (14, 2) and (15, 2)
Median: 7.5
We must keep track of the median for every split, because that is actually a vital piece of knowledge in this algorithm. They don't show it very clearly on the slides, but knowing the median value (where you split a set to make two sets) is essential to solving this question quickly.
We keep track of a value they call delta in the algorithm. Ugh I don't know why most computer scientists absolutely suck at naming variables, you need to have descriptive names when you code so you don't forget what the f000 you coded 10 years ago, so instead of delta let's call this value our-shortest-twig-from-the-median-so-far
Since we have the median value of 7.5 let's go and see what our-shortest-twig-from-the-median-so-far is for Set1 and Set2, respectively:
Set1 : shortest-twig-from-the-median-so-far 2.5 (5 to m where m is 7.5)
Set 2: shortest-twig-from-the-median-so-far 3.5 (looking at 11 to m)
So I think the key take-away from the algorithm is that this shortest-twig-from-the-median-so-far is something that you're trying to improve upon every time you divide a set.
Since S1 in our case has 2 elements only, we are done with the left set, and we have 3 in the right set, so we continue dividing:
S2 = { (11,2) (14,2) (15,2) }
What do you do? You make a new median, call it S2-median
S2-median is halfway between 15 and 11... or 13, right? My math may be fuzzy, but I think that's right so far.
So let's look at the shortest-twig-so-far-for-our-right-side-with-median-thirteen ...
15 to 13 is... 2
11 to 13 is .... 2
14 to 13 is ... 1 (!!!)
So our m value or shortest-twig-from-the-median-so-far is improved (where we updated our median from before because we're in a new chunk or Set...)
Now that we've found it we know that (14, 2) is one of the points that satisfies the shortest pair equation. You can then check exhaustively against the points in this subset (15, 11, 14) to see which one is the closer one.
Clearly, (15,2) and (14,2) are the winning pair in this case.
Does that make sense? You must keep track of the median when you cut the set, and keep a new median for everytime you cut the set until you have only 2 elements remaining on each side (or in our case 3)
The magic is in the median or shortest-twig-from-the-median-so-far
Thanks for asking this question, I went in not knowing how this algorithm worked but found the right highlighted bullet point on the slide and rolled with it. Do you get it now? I don't know how to explain the median magic other than binary search is f000ing awesome.
I am trying to implement a stochastic ant colony optimisation algorithm, and I'm having trouble working out how to implement movement choices based on probabilities.
the standard (greedy) version that I have implemented so far is that an ant m at a vertex i on a graph G = (V,E) where E is the set of edges (i, j), will choose the next vertex j based on the following criteria:
j = argmax(<fitness function for j>)
such that j is connected to i
the problem I am having is in trying to implement a stochastic version of this, so that now the criteria for choosing a new vertex, j is:
P(j) = <fitness function for j>/sum(<fitness function for J>)
where P(j) is the probability of choosing vertex j,
such j is connected to i,
and J is the set of all vertices connected to i
I understand the mathematics behind it, I am just having trouble working out how i should actually implement it.
if, say, i have 3 vertices connected to i, each with a probability of 0.2, 0.3, 0.5 - what is the best way to make the selection? should I just randomly select a vertex j, then generate a random number r in the range (0,1) and if r >= P(j), select vertex j? or is there a better way?
Looking at the problem statement, I think you are not trying to visit all nodes (connected to i (say) ), but some of the nodes based on some probability distribution. Lets take an example:
You have a node i and connected to it are 5 nodes, a1...a5, with probabilities p1...p5, such that sum(p_i) = 1. No, say the precision of probabilities that you consider is 2 places after decimal. Also, you dont want to visit all 5 nodes, but only k of them. Lets say, in this example, k = 2. So, since 2 places of decimal is your probability precision, add 3 to it to increase normality of probability distribution in the random function. (You can change this 3 to any number of your choice, as far as performance is concerned) (Since you have not tagged any language, I'll take example of java's nextInt() function to generate random numbers.)
Lets give some values:
p1...p5 = {0.17, 0.11, 0.45, 0.03, 0.24}
Now, in a loop from 1 to k, generate a random number from (0...10^5). {5 = 2 + 3, ie. precision + 3}. If the generated number is from 0 to 16999, go with node a1, 17000 to 27999, go with a2, 28000 to 72999, go with a3...and so on. You get the idea.
What you're trying to implement is a weighted random choice depending on the probabilities for the components of the solution, or a random proportional selection rule on ACO terms. Here is an snippet of the implementation of this rule on the Isula Framework:
double value = random.nextDouble();
while (componentWithProbabilitiesIterator.hasNext()) {
Map.Entry<C, Double> componentWithProbability = componentWithProbabilitiesIterator
.next();
Double probability = componentWithProbability.getValue();
total += probability;
if (total >= value) {
nextNode = componentWithProbability.getKey();
getAnt().visitNode(nextNode);
return true;
}
}
You just need to generate a random value between 0 and 1 (stored in value), and start accumulating the probabilities of the components (on the total variable). When the total exceeds the threshold defined in value, we have found the component to add to the solution.
I have a collection of objects, each of which has a weight and a value. I want to pick the pair of objects with the highest total value subject to the restriction that their combined weight does not exceed some threshold. Additionally, I am given two arrays, one containing the objects sorted by weight and one containing the objects sorted by value.
I know how to do it in O(n2) but how can I do it in O(n)?
This is a combinatorial optimization problem, and the fact the values are sorted means you can easily try a branch and bound approach.
I think that I have a solution that works in O(n log n) time and O(n) extra space. This isn't quite the O(n) solution you wanted, but it's still better than the naive quadratic solution.
The intuition behind the algorithm is that we want to be able to efficiently determine, for any amount of weight, the maximum value we can get with a single item that uses at most that much weight. If we can do this, we have a simple algorithm for solving the problem: iterate across the array of elements sorted by value. For each element, see how much additional value we could get by pairing a single element with it (using the values we precomputed), then find which of these pairs is maximum. If we can do the preprocessing in O(n log n) time and can answer each of the above queries in O(log n) time, then the total time for the second step will be O(n log n) and we have our answer.
An important observation we need to do the preprocessing step is as follows. Our goal is to build up a structure that can answer the question "which element with weight less than x has maximum value?" Let's think about how we might do this by adding one element at a time. If we have an element (value, weight) and the structure is empty, then we want to say that the maximum value we can get using weight at most "weight" is "value". This means that everything in the range [0, max_weight - weight) should be set to value. Otherwise, suppose that the structure isn't empty when we try adding in (value, weight). In that case, we want to say that any portion of the range [0, weight) whose value is less than value should be replaced by value.
The problem here is that when we do these insertions, there might be, on iteration k, O(k) different subranges that need to be updated, leading to an O(n2) algorithm. However, we can use a very clever trick to avoid this. Suppose that we insert all of the elements into this data structure in descending order of value. In that case, when we add in (value, weight), because we add the elements in descending order of value, each existing value in the data structure must be higher than our value. This means that if the range [0, weight) intersects any range at all, those ranges will automatically be higher than value and so we don't need to update them. If we combine this with the fact that each range we add always spans from zero to some value, the only portion of the new range that could ever be added to the data structure is the range [weight, x), where x is the highest weight stored in the data structure so far.
To summarize, assuming that we visit the (value, weight) pairs in descending order of value, we can update our data structure as follows:
If the structure is empty, record that the range [0, value) has value "value."
Otherwise, if the highest weight recorded in the structure is greater than weight, skip this element.
Otherwise, if the highest weight recorded so far is x, record that the range [weight, x) has value "value."
Notice that this means that we are always splitting ranges at the front of the list of ranges we have encountered so far. Because of this, we can think about storing the list of ranges as a simple array, where each array element tracks the upper endpoint of some range and the value assigned to that range. For example, we might track the ranges [0, 3), [3, 9), and [9, 12) as the array
3, 9, 12
If we then needed to split the range [0, 3) into [0, 1) and [1, 3), we could do so by prepending 1 to he list:
1, 3, 9, 12
If we represent this array in reverse (actually storing the ranges from high to low instead of low to high), this step of creating the array runs in O(n) time because at each point we just do O(1) work to decide whether or not to add another element onto the end of the array.
Once we have the ranges stored like this, to determine which of the ranges a particular weight falls into, we can just use a binary search to find the largest element smaller than that weight. For example, to look up 6 in the above array we'd do a binary search to find 3.
Finally, once we have this data structure built up, we can just look at each of the objects one at a time. For each element, we see how much weight is left, use a binary search in the other structure to see what element it should be paired with to maximize the total value, and then find the maximum attainable value.
Let's trace through an example. Given maximum allowable weight 10 and the objects
Weight | Value
------+------
2 | 3
6 | 5
4 | 7
7 | 8
Let's see what the algorithm does. First, we need to build up our auxiliary structure for the ranges. We look at the objects in descending order of value, starting with the object of weight 7 and value 8. This means that if we ever have at least seven units of weight left, we can get 8 value. Our array now looks like this:
Weight: 7
Value: 8
Next, we look at the object of weight 4 and value 7. This means that with four or more units of weight left, we can get value 7:
Weight: 7 4
Value: 8 7
Repeating this for the next item (weight six, value five) does not change the array, since if the object has weight six, if we ever had six or more units of free space left, we would never choose this; we'd always take the seven-value item of weight four. We can tell this since there is already an object in the table whose range includes remaining weight four.
Finally, we look at the last item (value 3, weight 2). This means that if we ever have weight two or more free, we could get 3 units of value. The final array now looks like this:
Weight: 7 4 2
Value: 8 7 3
Finally, we just look at the objects in any order to see what the best option is. When looking at the object of weight 2 and value 3, since the maximum allowed weight is 10, we need tom see how much value we can get with at most 10 - 2 = 8 weight. A binary search over the array tells us that this value is 8, so one option would give us 11 weight. If we look at the object of weight 6 and value 5, a binary search tells us that with five remaining weight the best we can do would be to get 7 units of value, for a total of 12 value. Repeating this on the next two entries doesn't turn up anything new, so the optimum value found has value 12, which is indeed the correct answer.
Hope this helps!
Here is an O(n) time, O(1) space solution.
Let's call an object x better than an object y if and only if (x is no heavier than y) and (x is no less valuable) and (x is lighter or more valuable). Call an object x first-choice if no object is better than x. There exists an optimal solution consisting either of two first-choice objects, or a first-choice object x and an object y such that only x is better than y.
The main tool is to be able to iterate the first-choice objects from lightest to heaviest (= least valuable to most valuable) and from most valuable to least valuable (= heaviest to lightest). The iterator state is an index into the objects by weight (resp. value) and a max value (resp. min weight) so far.
Each of the following steps is O(n).
During a scan, whenever we encounter an object that is not first-choice, we know an object that's better than it. Scan once and consider these pairs of objects.
For each first-choice object from lightest to heaviest, determine the heaviest first-choice object that it can be paired with, and consider the pair. (All lighter objects are less valuable.) Since the latter object becomes lighter over time, each iteration of the loop is amortized O(1). (See also searching in a matrix whose rows and columns are sorted.)
Code for the unbelievers. Not heavily tested.
from collections import namedtuple
from operator import attrgetter
Item = namedtuple('Item', ('weight', 'value'))
sentinel = Item(float('inf'), float('-inf'))
def firstchoicefrombyweight(byweight):
bestsofar = sentinel
for x in byweight:
if x.value > bestsofar.value:
bestsofar = x
yield (x, bestsofar)
def firstchoicefrombyvalue(byvalue):
bestsofar = sentinel
for x in byvalue:
if x.weight < bestsofar.weight:
bestsofar = x
yield x
def optimize(items, maxweight):
byweight = sorted(items, key=attrgetter('weight'))
byvalue = sorted(items, key=attrgetter('value'), reverse=True)
maxvalue = float('-inf')
try:
i = firstchoicefrombyvalue(byvalue)
y = i.next()
for x, z in firstchoicefrombyweight(byweight):
if z is not x and x.weight + z.weight <= maxweight:
maxvalue = max(maxvalue, x.value + z.value)
while x.weight + y.weight > maxweight:
y = i.next()
if y is x:
break
maxvalue = max(maxvalue, x.value + y.value)
except StopIteration:
pass
return maxvalue
items = [Item(1, 1), Item(2, 2), Item(3, 5), Item(3, 7), Item(5, 8)]
for maxweight in xrange(3, 10):
print maxweight, optimize(items, maxweight)
This is similar to Knapsack problem. I will use naming from it (num - weight, val - value).
The essential part:
Start with a = 0 and b = n-1. Assuming 0 is the index of heaviest object and n-1 is the index of lightest object.
Increase a til objects a and b satisfy the limit.
Compare current solution with best solution.
Decrease b by one.
Go to 2.
Update:
It's the knapsack problem, except there is a limit of 2 items. You basically need to decide how much space you want for the first object and how much for the other. There is n significant ways to split available space, so the complexity is O(n). Picking the most valuable objects to fit in those spaces can be done without additional cost.
I would like to learn how to select weighted items. For example : I want to fetch questions from a pool but if some one can't give right answer to a question, that causes this question to double its weight and increase the probability of being selected again later on.
Have a class which keeps the item:weight pairs (key=item:value=weight) in a hash table.
The class should also maintain a total_weight variable, which is the sum of all the weights in the hash table. The class' methods to add_item, remove_item, and update_weight for an item should keep the total_weight updated. This avoids having to recalculate the total for every choice.
To choose an item:
Use a random number such that 1<=random_number<=total_weight.
Iterate over the item:weight pairs in the hash table, summing the weights until the random number is <= that running sum. When that happens, the key of the pair you're on is the chosen item.
This is like rolling an imaginary die whose size is the sum of all the weights. For every roll, each item has its own range of numbers on the die, with the size of each range equal to its item's weight. If the roll result falls within an item's range, that item is the chosen one.
Editing to add the following sample code after the request in the comment below. Tested this with Python 2.5.2:
from random import randint # Import randint function from random module.
class WeightedCollection(object):
def __init__(self):
self.total_weight = 0
self.items = {} # This is a python dictionary == a hash table
def add_item(self, item, weight):
self.items[item] = weight
self.total_weight += weight
def remove_item(self, item):
self.total_weight -= self.items[item] # Subtracts the weight.
del(self.items[item])
def update_weight(self, item, new_weight):
self.total_weight += (new_weight - self.items[item])
self.items[item] = new_weight
def get_random_item(self):
''' Returns random selection but weighted by item weights. '''
# Result of call below is 1 <= random_number <= self.total_weight...
random_number = randint(1, self.total_weight)
sum_so_far = 0
# For every item and its weight...
for item, weight in self.items.iteritems():
sum_so_far += weight
if random_number <= sum_so_far:
return item
# Usage demo...
questions = WeightedCollection()
questions.add_item('What is your name?', 1)
questions.add_item('What is your favorite color?', 50)
questions.add_item('What is the meaning to life?', 100)
print 'Here is what the dictionary looks like:'
print questions.items
print ''
print "Total weight:", questions.total_weight
print ''
print 'Some sample random picks...'
for i in range(5):
print questions.get_random_item()
And here is the output:
Here is what the dictionary looks like:
{'What is the meaning to life?': 100, 'What is your name?': 1, 'What is your favorite color?': 50}
Total weight: 151
Some sample random picks...
What is your favorite color?
What is the meaning to life?
What is the meaning to life?
What is your favorite color?
What is the meaning to life?
Keep around an array of the candidate items. If one item has weight 2, put it in the array twice, generally if one has weight n put it in there n times. Then select a random element from the array. Ta-daaa.
I like #André Hoffmann's idea of using a binary tree, in which every leaf node corresponds to a question, and every intermediate node stores the sum of the weight of its child nodes. But he says the tree needs to be re-created every time a weight changes.
Actually, this need not be the case! When you change the weight of a given leaf, you only need to update the weights of those nodes between it and the root of the tree. But...you also need some way to find the node within the tree, if you want to modify it.
So my suggestion is to use a self-balancing binary tree (e.g. a red-black tree, AVL tree, etc), which is ordered by the question ID. Operations on the tree need to maintain the property that the weight of any node is equal to the sum of the weights of its children.
With this data structure, the root node's weight W is equal to the sum of the weights of all the questions. You can retrieve a question either by question ID, or by a random weight (between zero and W). This operation, as well as insertions, deletions, or updating the weight of a question are all O(log n).
Have a look at this this(scroll down for the code).
EDIT for the critics:)
The code on this thread I linked shows how to implement a binary tree approach that actually works with weights and doesn't store masses of elements in an array to achieve a weighted probability. Then again, it's pretty inefficient when the weights change very often as the binary tree has to be re-created every time a weight changes.
EDIT2:
See Todd Owen's post about using self-balancing trees. The tree obviously does not have to be re-created every time a weight changes. That part just isn't included in the implementation I linked and needs to be added if your weights change a lot.