Randomised Enumeration of Small Witnesses Using a Decision Oracle
 209 Downloads
Abstract
Many combinatorial problems involve determining whether a universe of n elements contains a witness consisting of k elements which have some specified property. In this paper we investigate the relationship between the decision and enumeration versions of such problems: efficient methods are known for transforming a decision algorithm into a search procedure that finds a single witness, but even finding a second witness is not so straightforward in general. We show that, if the decision version of the problem can be solved in time \(f(k) \cdot poly(n)\), there is a randomised algorithm which enumerates all witnesses in time \(e^{k + o(k)} \cdot f(k) \cdot poly(n) \cdot N\), where N is the total number of witnesses. If the decision version of the problem is solved by a randomised algorithm which may return false negatives, then the same method allows us to output a list of witnesses in which any given witness will be included with high probability. The enumeration algorithm also gives rise to an efficient algorithm to count the total number of witnesses when this number is small.
Keywords
Enumeration algorithms Parameterized complexity Randomized algorithms Approximate enumeration Color coding1 Introduction
Many wellknown combinatorial decision problems involve determining whether a universe U of n elements contains a witness W consisting of exactlyk elements which have some specified property; we refer to such problems as kwitness problems. Although the decision problems themselves are of interest, it is often not sufficient for applications to output simply “yes” or “no”: we need to find one or more witnesses. The issue of finding a single witness using an oracle for the decision problem has previously been investigated by Björklund et al. [6], motivated by the fact that the fastest known parameterised algorithms for a number of widely studied problems (such as graph motif [5] and kpath [4]) are nonconstructive in nature. Moreover, for some problems (such as kClique or Independent Set [3] and kEven Subgraph [17]) the only known FPT decision algorithm relies on a Ramsey theoretic argument which says the answer must be “yes” provided that the input graph avoids certain easily recognisable structures.

INCORA(X, U, k)

Input: \(X \subseteq U\) and \(k \in \mathbb {N}\)

Output: 1 if some witness of size k in U is entirely contained in X; 0 otherwise.
Given access to an oracle of this kind, a naïve approach easily finds a single witness using \(\Theta (n)\) calls to INCORA: we successively delete elements of the universe, following each deletion with an oracle call, and if the oracle answers “no” we reinsert the last deleted element and continue. Assuming we start with a yesinstance, this process will terminate when only k elements remain, and these k elements must form a witness. In [6], ideas from combinatorial group testing are used to make a substantial improvement on this strategy for the extraction of a single witness: rather than deleting a single element at a time, large subsets are discarded (if possible) at each stage. This gives an algorithm that extracts a witness with only \(2k\left( \log _2\left( \frac{n}{k}\right) +2\right) \) oracle queries.
However, neither of these approaches for finding a single witness can immediately be extended to find all witnesses, a problem which is of interest even if an efficient decision algorithm does output a single witness. Both approaches for finding a first witness rely on the fact that we can safely delete some subset of elements from our universe provided we know that what is left still contains at least one witness; if we need to look for a second witness, the knowledge that at least one witness will remain is no longer sufficient to guarantee we can delete a given subset. Of course, for any kwitness problem we can check all possible subsets of size k, and hence enumerate all witnesses, in time \(O(n^k)\); indeed, if every set of k vertices is in fact a witness then we will require this amount of time simply to list them all. However, we can seek to do much better than this when the number of witnesses is small by making use of a decision oracle.

EXTORA(X,Y,U,k)

Input: \(X \subseteq U\), \(Y \subseteq X\), and \(k \in \mathbb {N}\)

Output: 1 if there exists a witness W with \(Y \subseteq W \subseteq X\); 0 otherwise.
Note that, with only the inclusion oracle, we can determine whether there is a witness that does not contain some element x (we simply call INCORA(\(U \setminus \{x\}\), U, k)), but we cannot determine whether there is a witness which does contain x. Moreover, as we will show in Sect. 3, there are natural (selfcontained) kwitness problems for which the inclusion problem can be solved efficiently but there is no fptalgorithm for the extension decision problem unless FPT=W[1]. This motivates the development of enumeration algorithms that do not rely on such an oracle.
The main result of this paper is just such an algorithm; specifically, we prove the following theorem.
Theorem 1.1
The key tool we use to obtain this algorithm is a colour coding method, using a family of kperfect hash functions. This technique was introduced by Alon et al. in [1] and has been widely used in the design of parameterised algorithms for decision and approximate counting (see for example [15, Chapters 13 and 14] and [12, Chapter 8]), but to the best of the author’s knowledge has not yet been applied to enumeration problems.
The main limitation of Theorem 1.1 is that it requires access to a deterministic inclusion oracle INCORA which always returns the correct answer. However, in a number of cases (including kPath [4] and Graph Motif [5]) the fastest known decision algorithm for a selfcontained kwitness problem (and hence for the corresponding inclusion problem) is in fact randomised and has a small probability of returning an incorrect answer. We will also show that the same algorithm can be used in this case, at the expense of a small increase in the expected running time (if the oracle can return false positives) and the loss of the guarantee that we will output every witness exactly once: for each witness in the instance, there is a small probability that we will omit it from the list due to the oracle returning false negatives. Specifically, we prove the following theorem.
Theorem 1.2
This result initiates the study of approximate algorithms for enumeration problems: in contrast with the wellestablished field of approximate counting, this relaxation of the requirements for enumeration does not seem to have been addressed in the literature to date.
In the study of counting complexity it is standard practice, when faced with a \(\#\mathsf{{P}}\)hard problem, to investigate whether there is an efficient method to solve the counting problem approximately. The answer to this question is considered to be “yes” if and only if the problem admits a fully polynomial randomised approximation scheme (FPRAS), defined as follows.
Definition
In the parameterised setting, the analogue of this is a fixed parameter tractable randomised approximation scheme (FPTRAS), in which the running time is additionally allowed to depend arbitrarily on the parameter.
Perhaps the most obvious way to translate this notion in to the setting of enumeration would be to look for an algorithm which, with probability at least \((1  \delta )\), would output at least \((1\epsilon )\)proportion of all witnesses. In the setting of counting, all witnesses are essentially interchangeable, so it makes sense to consider only the total number of objects counted in relation to the true answer. However, this definition perhaps allows too much freedom in the setting of enumeration: we could design an algorithm which satisfies these requirements and yet will never output some collection of hardtofind witnesses, so long as this collection is not too large compared with the total number of witnesses.
Instead, we propose here a more demanding notion of approximate enumeration: given \(\epsilon > 0\), we want a (randomised) algorithm such that, for any witness W, the probability we output W is at least \(1\epsilon \). This implies that we will, with high probability (depending on \(\epsilon \)) output a large proportion of all possible witnesses, but also ensures that we cannot choose to ignore certain potential witnesses altogether. It may also be desirable to permit a witness to be repeated in the output with small probability: we can allow this flexibility by requiring only that, for each witness W, the probability that W is included in the output exactly once is at least \(1  \epsilon \). We give a formal definition of efficient approximate enumeration in Sect. 2.
Theorem 1.1 is proved in Sect. 4, and Theorem 1.2 in Sect. 5. We then discuss some implications of our enumeration algorithms for the complexity of related counting problems in Sect. 6. We begin in Sect. 2 with some background on relevant complexity theoretic notions, before discussing the hardness of the extension version of some selfcontained kwitness problems in Sect. 3.
2 Parameterised Enumeration
There are two natural measures of the size of a selfcontained kwitness problem, namely the number of elements n in the universe and the number of elements k in each witness, so the running time of algorithms is most naturally discussed in the setting of parameterised complexity. There are two main complexity issues to consider in the present setting: first of all, as usual, the running time, and secondly the number of oracle calls required.

an expectedtotalfpt algorithm enumerates all solutions and terminates in expected time \(f(k)\cdot n^{O(1)}\);

an expecteddelayfpt algorithm enumerates all solutions with expected delay at most \(f(k) \cdot n^{O(1)}\) between the times at which one solution and the next are output (and the same bound applies to the time before outputting the first solution, and between outputting the final solution and terminating);

an expectedincrementalfpt algorithm enumerates all solutions with expected delay at most \(f(k) \cdot (n + i)^{O(1)}\) between outputting the ith and \((i+1)\)th solution;

an expectedoutputfpt algorithm enumerates all solutions and terminates in expected time \(f(k) \cdot (n+N)^{O(1)}\), where N is the total number of solutions enumerated.
In the setting of approximate enumeration, we define a fully output polynomial randomised enumeration scheme (FOPRES) to be an algorithm which, given an instance I of an enumeration problem (with total input size n) and a rational \(\epsilon \in (0,1)\), outputs, in time bounded by a polynomial function of n, N and \(\epsilon ^{1}\) (where N is the total number of solutions to I), a list of solutions to I with the property that, for any solution W, the probability that W appears exactly once in the list is at least \(1  \epsilon \). In the parameterised setting, we analogously define a fully output fpt randomised enumeration scheme (FOFPTRES) to be an algorithm which, given an instance I of a parameterised enumeration problem (with total input size n and parameter k) and a rational \(\epsilon \in (0,1)\), outputs, in time bounded by \(f(k)\cdot p(n,N,\epsilon ^{1})\), where p is a polynomial, f is any computable function, and N is the total number of solutions to I, a list of solutions to I with the property that, for any solution W, the probability that W appears exactly once in the list is at least \(1  \epsilon \). An expectedFOPRES (respectively expectedFOFPTRES) is a randomised algorithm which satisfies the definition of a FOPRES (resp. FOFPTRES) if we replace the condition on the running time by the same condition on the expected running time. We can make analogous definitions for totalpolynomial, totalfpt, delaypolynomial etc.
Under these definitions, Theorem 1.2 says that, if there is a randomised FPT decision algorithm for the inclusion version of a kwitness problem with error probability less than a half, then the corresponding enumeration problem admits a FOFPTRES.
3 Hardness of the Extension Problem
Many combinatorial problems have a very useful property, often referred to as selfreducibility, which allows a search or enumeration problem to be reduced to (smaller instances of) the corresponding decision problem in a very natural way (see [9, 20, 27]). A problem is selfreducible in this sense if the existence of an efficient decision procedure (answering the question: “Does the universe contain at least one witness of size k?”) implies that there is an efficient algorithm to solve the extension decision problem (equivalent to EXTORA). While many selfcontained kwitness problems do have this property, we will demonstrate that there exist selfcontained kwitness problems that do not (unless FPT=W[1]), and so an enumeration procedure that makes use only of INCORA and not EXTORA is desirable.
In order to demonstrate this, we show that there exist selfcontained kwitness problems whose decision versions belong to FPT, but for which the corresponding extension decision problem is W[1]hard. We will consider the following problem, which is clearly a selfcontained kwitness problem.
kClique or Independent Set
Input: A graph \(G = (V,E)\) and \(k \in \mathbb {N}\).
Parameter: k.
Question: Is there a kvertex subset of V that induces either a clique or an independent set?
This problem is known to belong to FPT [3]: any graph with at least \(2^{2k}\) vertices must be a yesinstance by Ramsey’s Theorem. We now turn our attention to the extension version of the problem, defined as follows.
kExtension Clique or Independent Set
Input: A graph \(G = (V,E)\), a subset \(U \subseteq V\) and \(k \in \mathbb {N}\).
Parameter: k.
Question: Is there a kvertex subset S of V, with \(U \subseteq S\), that induces either a clique or an independent set?
It is straightforward to adapt the hardness proof for kMulticolour Clique or Independent Set [22, Proposition 3.7] to show that kExtension Clique or Independent Set is W[1]hard.
Proposition 3.1
kExtension Clique or Independent Set is W[1]hard.
Proof
We prove this result by means of a reduction from the W[1]complete problem kClique. Let (G, k) be the input to an instance of kClique. We now define a new graph \(G'\), obtained from G by adding one new vertex v, and an edge from v to every vertex \(u \in V(G)\). It is then straightforward to verify that \((G',\{v\},k+1)\) is a yesinstance for kExtension Clique or Independent Set if and only if G contains a clique of size k. \(\square \)
This demonstrates that kExtension Clique or Independent Set is a problem for which there exists an efficient decision procedure but no efficient algorithm for the extension version of the decision problem (unless FPT=W[1]). Both of these arguments (inclusion of the decision problem in FPT, and hardness of the extension version) can easily be adapted to demonstrate that the following problem exhibits the same behaviour.
kInduced Regular Subgraph
Input: A graph \(G = (V,E)\) and \(k \in \mathbb {N}\).
Parameter: k.
Question: Is there a kvertex subset of V that induces a subgraph in which every vertex has the same degree?
Indeed, the same method can be applied to any problem in which putting a restriction on the degree of one of the vertices in the witness guarantees that the witness induces a clique (or some other induced subgraph for which it is W[1]hard to decide inclusion in an arbitrary input graph).
4 The Randomised Enumeration Algorithm
In this section we describe our randomised witness enumeration algorithm and analyse its performance when used with a deterministic oracle, thus proving Theorem 1.1.
As mentioned above, our algorithm relies on a colour coding technique. A family \(\mathcal {F}\) of hash functions from [n] to [k] is said to be kperfect if, for every subset \(A \subset [n]\) of size k, there exists \(f \in \mathcal {F}\) such that the restriction of f to A is injective. We will use the following bound on the size of such a family of hash functions.
Theorem 4.1
[24] For all \(n, k \in \mathbb {N}\) there is a kperfect family \(\mathcal {F}_{n,k}\) of hash functions from [n] to [k] of cardinality \(e^{k + o(k)} \cdot \log n\). Furthermore, given n and k, a representation of the family \(\mathcal {F}_{n,k}\) can be computed in time \(e^{k + o(k)} \cdot n \log n\).
Our strategy is to solve a collection of \(e^{k + o(k)} \cdot \log n\)colourful enumeration problems, one corresponding to each element of a family \(\mathcal {F}\) of kperfect hash functions. In each of these problems, our goal is to enumerate all witnesses that are colourful with respect to the relevant element f of \(\mathcal {F}\) (those in which each element is assigned a distinct colour by f). Of course, we may discover the same witness more than once if it is colourful with respect to two distinct elements in \(\mathcal {F}\), but it is straightforward to check for repeats of this kind and omit duplicate witnesses from the output. It is essential in the algorithm that we use a deterministic construction of a kperfect family of hash functions rather than the randomised construction also described in [1], as the latter method would allow the possibility of witnesses being omitted (with some small probability).
The advantage of solving a number of colourful enumeration problems is that we can split the problem into a number of subproblems with the only requirement being that we preserve witnesses in which every element has a different colour (rather than all witnesses). This makes it possible to construct a number of instances, each (roughly) half the size of the original instance, such that every colourful witness survives in at least one of the smaller instances. More specifically, for each kperfect hash function we explore a search tree: at each node, we split every colourclass randomly into (almost) equalsized parts, and then branch to consider each of the \(2^k\) combinations that includes one (nonempty) subset of each colour, provided that the union of these subsets still contains at least one witness (as determined by the decision oracle). This simple pruning of the search tree will not prevent us exploring “deadends” (where we pursue a particular branch due to the presence of a noncolourful witness), but turns out to be sufficient to make it unlikely that we explore very many branches that do not lead to colourful witnesses.
We describe the algorithm in pseudocode (Algorithm 1), making use of two subroutines. In addition to our oracle INCORA(X,U,k), we also define a procedure RANDPART(X) which we use, while exploring the search tree, to obtain a random partition of a subset of the universe.
RANDPART(X)
Input: \(X \subseteq U\)
We prove the correctness of the algorithm and discuss the space used in Sect. 4.1, and bound the expected running time in Sect. 4.2.
4.1 Correctness of the Algorithm
In order to prove that our algorithm does indeed output every witness exactly once, we begin by showing that we will identify a given kelement subset X during the iteration corresponding to the hashfunction \(f \in \mathcal {F}\) if and only if X is a colourful witness with respect to f.
Lemma 4.1
 1.
X is a witness, and
 2.
X is colourful with respect to f.
Proof
 1.
there is some witness W such that \(W \subseteq B\), and
 2.
B contains at least one vertex receiving each colour under f.
Conversely, suppose that \(W = \{w_1,\ldots ,w_k\}\) is a witness such that \(f(w_i) = i\) for each \(1 \le i \le k\); we need to show that we will at some stage execute lines 9–11 with \(A = W\). We argue that at the start of each execution of the while loop, if W has not yet been output, there must be some subset B in the queue such that \(W \subseteq B\). This invariant clearly holds before the first execution of the loop (U will have been inserted into Q, as U contains at least one witness W). Now suppose that the invariant holds before starting some execution of the while loop. Either we execute lines 9–11 with \(A = W\) on this iteration (in which case we are done), or else we proceed to line 13. Now, for \(1 \le i \le k\), set \(j_i\) to be either 1 or 2 in such a way that \(w_i \in A_i^{(j_i)}\). The subset \(A_{\mathbf {j}}\), where \(\mathbf {j} = (j_1,\ldots ,j_k)\) will then pass both tests for insertion into Q, and \(W \subseteq A_{\mathbf {j}}\) by construction, so the invariant holds when we exit the while loop. Since the algorithm only terminates when Q is empty, it follows that we must eventually execute lines 9–11 with \(A = W\). \(\square \)
The key property of kperfect families of hash functions then implies that the algorithm will identify every witness; it remains only to ensure that we avoid outputting any witness more than once. This is the purpose of lines 9–11 in the pseudocode. We know from Lemma 4.1 that we find a given witness W while considering the hashfunction f if and only if W is colourful with respect to f: thus, in order to determine whether we have found the witness in question before, it suffices to verify whether it is colourful with respect to any of the colourings previously considered. Hence we see that every witness is output exactly once, as required.
Note that the most obvious strategy for avoiding repeats would be to maintain a list of all the witnesses we have output so far, and check for membership of this list; however, in general there might be as many as \(\left( {\begin{array}{c}n\\ k\end{array}}\right) \) witnesses, so both storing this list and searching it would be costly. The approach used here means that we only have to store the family \(\mathcal {F}\) of kperfect hash functions (requiring space \(e^{k + o(k)} n \log n\)). Since each execution of the outer for loop clearly requires only polynomial space, the total space complexity of the algorithm is at most \(e^{k + o(k)} n^{O(1)}\), as required.
4.2 Expected Running Time
We know from Theorem 4.1 that a family \(\mathcal {F}\) of kperfect hash functions from U to [k], with \(\mathcal {F} = e^{k + o(k)}\log n\), can be computed in time \(e^{k + o(k)}n \log n\); thus line 2 can be executed in time \(e^{k + o(k)}n \log n\) and the total number of iterations of the outer forloop (lines 2–34) is at most \(e^{k + o(k)} \log n\).
Moreover, it is clear that each iteration of the while loop (lines 6–26) makes at most \(2^k\) oracle calls. If an oracle call can be executed in time \(g(k)\cdot n^{O(1)}\) for some computable function g, then the total time required to perform each iteration of the while loop is at most \(\max \{\mathcal {F},kn + 2^k \cdot g(k)\cdot n^{O(1)}\} = e^{k + o(k)}\cdot g(k) \cdot n^{O(1)}.\)
Thus it remains to bound the expected number of iterations of the while loop in any iteration of the outer forloop; we do this in the next lemma.
Lemma 4.2
The expected number of iterations of the whileloop in any given iteration of the outer forloop is at most \(N \left( 1 + \lceil \log n \rceil \right) \), where N is the total number of witnesses in the instance.
Proof
We fix an arbitrary \(f \in \mathcal {F}\), and for the remainder of the proof restrict our attention to the iteration of the outer forloop corresponding to f.
We can regard this iteration of the outer forloop as the exploration of a search tree, with each node of the search tree indexed by some subset of U. The root is indexed by U itself, and every node has up to \(2^k\) children, each child corresponding to a different way of selecting one of the two randomly constructed subsets for each colour. A node may have strictly fewer than \(2^k\) children, as we use the oracle to prune the search tree (line 20), omitting the exploration of branches indexed by a subset of U that does not contain any witness (colourful or otherwise). Note that the search tree defined in this way has depth at most \(\lceil \log n \rceil \): at each level, the size of each colourclass in the indexing subset is halved (up to integer rounding).
In this search tree model of the algorithm, each node of the search tree corresponds to an iteration of the whileloop, and vice versa. Thus, in order to bound the expected number of iterations of the whileloop, it suffices to bound the expected number of nodes in the search tree.
Our oraclebased pruning method means that we can associate with every node v of the search tree some representative witness \(W_v\) (not necessarily colourful), such that \(W_v\) is entirely contained in the subset of U which indexes v. (Note that the choice of representative witness for a given node need not be unique.) We know that in total there are N witnesses; our strategy is to bound the expected number of nodes, at each level of the search tree, for which any given witness can be the representative.
Observe first that, if W is in fact a colourful witness with respect to f, then \(X_{W,d} = 1\) for every d: given a node whose indexing set contains W, exactly one of its children will be indexed by a set that contains W. So we will assume from now on that W intersects precisely \(\ell \) colour classes, where \(\ell < k\).
5 Using a Randomised Oracle
In this section we show that the method described in Sect. 4 will in fact work almost as well if we only have access to a randomised decision oracle, thus proving Theorem 1.2. The randomised decision procedures in [4, 5] only have onesided errors, but for the sake of generality we consider the effect of both false positives and false negatives on our algorithm.
False positives and false negatives will affect the behaviour of the algorithm in different ways. If the decision procedure gives false positives then, provided we add a check immediately before outputting a supposed witness that it really is a witness, the algorithm is still sure to output every witness exactly once; however, we will potentially waste time exploring unfruitful branches of the search tree due to false positives, so the expected running time of the algorithm will increase. If, on the other hand, our algorithm outputs false negatives, then this will not increase the expected running time; however, in this case, we can no longer be sure that we will find every witness as false negatives might cause us to prune the search tree prematurely. We will show, however, that we can still enumerate approximately in this case.
Before turning our attention to the specific effects of false positives and false negatives on the algorithm, we observe that, provided our randomised oracle returns the correct answer with probability greater than a half, we can obtain a decision procedure with a much smaller failure probability by making repeated oracle calls. We make the standard assumption that the events corresponding to each oracle call returning an error are independent.
Lemma 5.1
 1.
the probability of obtaining a false positive is at most \(2^{k}\), and
 2.
the probability of obtaining a false negative is at most \(\frac{\epsilon }{\lceil \log n \rceil + 1}\)
Proof
Our procedure is as follows: we make t oracle calls (where t is a value to be determined later) and output whatever is the majority answer from these calls. We need to choose t large enough to ensure that the probability that the majority answer is incorrect is at most \(\delta :=\min \left\{ 2^{k}, \frac{\epsilon }{\lceil \log n \rceil + 1}\right\} \).
The probability that we obtain the correct answer from a given oracle call is at least c, so the number of correct answers we obtain out of t trials is bounded below by the random variable X, where X has distribution \({{\mathrm{Bin}}}(t,c)\). Thus \(\mathbb {E}[X] = tc\). We will return the correct answer so long as \(X > \frac{t}{2}\).
We now show that, if the probability that our oracle gives a false positive is sufficiently small, then such errors do not increase the expected running time of Algorithm 1 too much. Just as when bounding the expected running time in Sect. 4.2, it suffices to bound the expected number of iterations of the while loop corresponding to a specific colouring f in our family \(\mathcal {F}\) of hash functions.
Lemma 5.2
Suppose that the probability that the oracle returns a false positive is at most \(\min \left\{ 2^{k}, \frac{1}{\lceil \log n \rceil + 1} \right\} \). Then the expected number of iterations of the whileloop in any given iteration of the outer forloop is at most \(O(N \cdot \log ^2n)\), where N is the total number of witnesses in the instance.
Proof
We fix an arbitrary \(f \in \mathcal {F}\), and for the remainder of the proof we restrict our attention or the iteration of the outer forloop corresponding to f. As in the proof of Lemma 4.2, we can regard this iteration of the outer forloop as the exploration of a search tree, and it suffices to bound the expected number of nodes in the search tree.
We can associate with each node of the search tree some subset of the universe, and we prune the search tree in such a way that we only have a node corresponding to a subset A of the universe if a call to the oracle with input A has returned yes. This means that for the node corresponding to the set A, either there is some representative witness \(W \subseteq X\), or the oracle gave a false positive. We call a node good if it has some representative witness, and bad if it is the result of a false positive. We already bounded the expected number of good nodes in the proof of Lemma 4.2, so it remains to show that the expected number of bad nodes is not too large.
We will assume initially that there is at least one witness, and so the root of the search tree is a good node. Now consider a bad node v in the search tree; v must have some ancestor u in the search tree such that u is good (note that the root will always be such an ancestor in this case). Since the subset of the universe associated with any node is a subset of that associated with its parent, no bad vertex can have a good descendant. Thus, any path from the root to the bad node v must consist of a segment of good nodes followed by a segment of bad nodes; we can therefore associate with every bad node v a unique good node \(good(v)\) such that \(good(v)\) is the last good node on the path from the root to v. In order to bound the expected number of bad nodes in the tree, our strategy is to bound, for each good node u, the number of bad nodes v such that \(good(v) = u\).
As in the proof of Lemma 4.2, we will write \(X_{W,d}\) for the number of nodes at depth d for which W is the representative witness. For \(c > d\), we further define \(Y_{W,d,c}\) to be the number of bad nodes v such that v is at depth c, \(good(v)\) is at depth d, and W is the representative witness for \(good(v)\).
If there is in fact no witness in the instance, we know that there are no good nodes in the tree. Moreover, the expected number of bad nodes at depth 0 is at most \(1/\left( \lceil \log n \rceil + 1 \right) \) (the probability that the oracle returns a false positive). Since we have already argued that the expected number of bad children of any node is at most 1, it follows that the expected number of bad nodes at each level is at most \(1/\left( \lceil \log n \rceil + 1 \right) \), and so the total expected number of bad nodes is at most \(1/\left( \lceil \log n \rceil + 1 \right) \left( 1 + \lceil \log n \rceil \right) = 1\). \(\square \)
To complete the proof of Theorem 1.2, it remains to show that, so long as the probability that the oracle returns a false negative is sufficiently small, our algorithm will output any given witness with high probability.
Lemma 5.3
Fix \(\epsilon \in (0,1)\), and suppose that the probability that the oracle returns a false negative is at most \(\frac{\epsilon }{\lceil \log n \rceil + 1}\). Then, for any witness W, the probability that the algorithm does not output W is at most \(\epsilon \).
Proof
By construction of \(\mathcal {F}\), we know that there is some \(f \in \mathcal {F}\) such that W is colourful with respect to f. We wil now restrict our attention to the iteration of the outer forloop corresponding to f; it suffices to demonstrate that we will output W during this iteration with probability at least \(1  \epsilon \).
6 Application to Counting
There is a close relationship between the problems of counting and enumerating all witnesses in a kwitness problem, since any enumeration algorithm can very easily be adapted into an algorithm that counts the witnesses instead of listing them. However, in the case that the number N of witnesses is large, an enumeration algorithm necessarily takes time at least \(\Omega (N)\), whereas we might hope for much better if our goal is simply to determine the total number of witnesses.
The family of selfcontained kwitness problems studied here includes subgraph problems, whose parameterised complexity from the point of view of counting has been a rich topic for research in recent years [10, 11, 14, 17, 18, 19, 22]. Many such counting problems, including those whose decision problem belongs to FPT, are known to be #W[1]complete (see [15] for background on the theory of parameterised counting complexity). Positive results in this setting typically exploit structural properties of the graphs involved (e.g. small treewidth) to design (approximate) counting algorithms for inputs with these properties, avoiding any dependence on N [2, 3, 18].
In this section we demonstrate how our enumeration algorithms can be adapted to give efficient (randomised) algorithms to solve the counting version of a selfcontained kwitness problem whenever the total number of witnesses is small. This complements the fact that a simple random sampling algorithm can be used for approximate counting when the number of witnesses is very large [22, Lemma 3.4], although there remain many situations which are not covered by either result.
We begin with the case in which we assume access to a deterministic oracle for the decision problem.
Theorem 6.1
 1.
if the number of witnesses in the instance of \(\Pi \) is at most M, outputs with probability at least \(1  \delta \) the exact number of witnesses in the instance;
 2.
if the number of witnesses in the instance of \(\Pi \) is strictly greater than M, always outputs “More than M.”
Proof
Note that Algorithm 1 can very easily be adapted to give a randomised counting algorithm which runs in the same time as the enumeration algorithm but, instead of listing all witnesses, simply outputs the total number of witnesses when it terminates. We may compute explicitly the expected running time of our randomised enumeration algorithm (and hence its adaptation to a counting algorithm) for a given selfcontained kwitness problem \(\Pi \) in terms of n, k and the total number of witnesses, N. We will write \(T(\Pi ,n,k,N)\) for this expected running time.
Now consider an algorithm A, in which we run our randomised counting algorithm for at most \(2T(\Pi ,n,k,M)\) steps; if the algorithm has terminated within this many steps, A outputs the value returned, otherwise A outputs “FAIL”. Since our randomised counting algorithm is always correct (but may take much longer than the expected time), we know that if A outputs a numerical value then this is precisely the number of witnesses in our problem instance. If the number of witnesses is in fact at most M, then the expected running time of the randomised counting algorithm is bounded by \(T(\Pi ,n,k,M)\), so by Markov’s inequality the probability that it terminates within \(2T(\Pi ,n,k,M)\) steps is at least 1 / 2. Thus, if we run A on an instance in which the number of witnesses is at most M, it will output the exact number of witnesses with probability at least 1 / 2.
To obtain the desired probability of outputting the correct answer, we repeat A a total of \(\lceil \log (\delta ^{1})\rceil \) times. If any of these executions of A terminates with a numerical answer that is at most M, we output this answer (which must be the exact number of witnesses by the argument above); otherwise we output “More than M.”
If the total number of witnesses is in fact less than or equal to M, we will output the exact number of witnesses unless A outputs “FAIL” every time it is run. Since in this case A outputs “FAIL” independently with probability at most 1 / 2 each time we run it, the probability that we output “FAIL” on every one of the \(\lceil \log (\delta ^{1})\rceil \) repetitions is at most \((1/2)^{\lceil \log (\delta ^{1})\rceil } \le 2^{\log \delta } = \delta \). Finally, note that if the number of witnesses is strictly greater than M, we will always output “More than M” since every execution of A must in this case return either “FAIL” or a numerical answer greater than M.
The total running time is at most \(O\left( \log (\delta ^{1}) \cdot T(\Pi ,n,k,M)\right) \) and hence, using the bound on the running time of our enumeration algorithm from Theorem 1.1, is bounded by \(e^{k + o(k)}\cdot g(k) \cdot n^{O(1)} \cdot M \cdot \log (\delta ^{1})\), as required. \(\square \)
Finally, we prove an analogous result in the case that we only have access to a randomised oracle.
Theorem 6.2
 1.
if \(N \le M\), outputs a number \(N'\) such that \((1\epsilon )N \le N' \le N\);
 2.
if \(N \ge M\), outputs either a number \(N'\) such that \((1  \epsilon ) N \le N' \le M\) or “More than M.”
Proof
 (a)
a number \(N'\) such that \((1  \epsilon ) N \le N' \le N\) if \(N \le M\), and
 (b)
either a number \(N'\) such that \((1\epsilon )N \le N' \le N\) or “FAIL” if \(N > M\).
To obtain the required procedure, we modify the enumeration algorithm used to prove Theorem 1.2 so that it counts the total number of witnesses found rather than listing them; we will run this randomised enumeration procedure with error probability \(\epsilon ^2/4\). We can compute explicitly the expected running time of this adapted algorithm for a given kwitness problem \(\Pi \) in terms of n, k, N and \(\epsilon \); we write \(T(\Pi , n, k, \epsilon , N)\) for this expectation. We will allow the adapted algorithm to run for time \(\displaystyle 4 T(\Pi ,n,k,\epsilon ,M)\), outputting “FAIL” if we have not terminated within this time.
There are two ways in which the procedure could fail to meet conditions (a) and (b). First of all, the adapted enumeration algorithm might not terminate within the required time. Secondly, it might terminate but with an answer \(N'\) where \(N' < (1\epsilon ) N\) (recall that enumeration algorithm never repeats a witness, and that we can verify each witness deterministically, ensuring that only ever output a subset of the witnesses actually present in the instance). In the remainder of the proof, we will argue that the probability of each of these two outcomes is strictly less than \(\frac{1}{4}\), so the probability of avoiding both is greater than \(\frac{1}{2}\), as required.
First, we bound the probability that the algorithm does not terminate within the required time. By Markov’s inequality, the probability that a random variable takes a value greater than four times its expectation is less than \(\frac{1}{4}\), so we see immediately that if the total number of witnesses is at most M then the probability that the algorithm fails to terminate within the permitted time is less than \(\frac{1}{4}\).
7 Conclusions and Open Problems
Many wellknown combinatorial problems satisfy the definition of the kwitness problems considered in this paper. We have shown that, given access to a deterministic decision oracle for the inclusion version of a kwitness problem (answering the question “does this subset of the universe contain at least one witness?”), there is a randomised algorithm which is guaranteed to enumerate all witnesses and whose expected number of oracle calls is at most \(e^{k + o(k)} \log ^2 n \cdot N\), where N is the total number of witnesses. Moreover, if the decision problem belongs to FPT (as is the case for many selfcontained kwitness problems), our enumeration algorithm is an expectedoutputfpt algorithm.
We have also shown that, in the presence of only a randomised decision oracle, we can use the same strategy to produce a list of witnesses so that the probability of any given witness appearing in the list is at least \(1  \epsilon \), with only a factor \(\log n\) increase in the expected running time. This result initiates the study of algorithms for approximate enumeration.
Our results also has implications for counting the number of witnesses. In particular, if the total number of witnesses is small (at most \(f(k) \cdot n^{O(1)}\) for some computable function f) then our enumeration algorithms can easily be adapted to give fptalgorithms that will, with high probability, calculate a good approximation to the number of witnesses in an instance of a selfcontained kwitness problem (in the setting where we have a deterministic decision oracle, we in fact obtain the exact answer with high probability). The resulting counting algorithms satisfy the conditions for a FPTRAS (Fixed Parameter Tractable Randomised Approximation Scheme, as defined in [3]), and in the setting with a deterministic oracle we do not even need the full flexibility that this definition allows: with probability \(1  \delta \) we will output the exact number of witnesses, rather than just an answer that is within a factor of \(1 \pm \epsilon \) of this quantity.
While the enumeration problem can be solved in a more straightforward fashion for selfcontained kwitness problems that have certain additional properties, we demonstrated that several selfcontained kwitness problems do not have these properties, unless FPT = W[1]. A natural line of enquiry arising from this work would be the characterisation of those selfcontained kwitness problems that do have the additional properties, namely those for which an fptalgorithm for the decision version gives rise to an fptalgorithm for the extension version of the decision problem.
Our approach assumed the existence of an oracle to determine whether a given subset of the universe contains a witness of size exactlyk. An interesting direction for future work would be to explore the extent to which the same techniques can be used if we only have access to a decision procedure that tells us whether some subset of the universe contains a witness of size at mostk.
Another key question that remains open after this work is whether the existence of an fptalgorithm for the inclusion version of a kwitness problem is sufficient to guarantee the existence of an (expected)delayfpt or (expected)incrementalfpt algorithm for the enumeration problem. Finally, it would be interesting to investigate whether the randomised algorithm given here can be derandomised.
Footnotes
References
 1.Alon, N., Yuster, R., Zwick, U.: Colorcoding. J. ACM 42(4), 844–856 (1995)MathSciNetCrossRefzbMATHGoogle Scholar
 2.Alon, N., Dao, P., Hajirasouliha, I., Hormozdiari, F., Sahinalp, S.C.: Biomolecular network motif counting and discovery by color coding. Bioinformatics 24(13), 241–249 (2008)CrossRefGoogle Scholar
 3.Arvind, V., Raman, V.: Approximation algorithms for some parameterized counting problems, ISAAC 2002, LNCS, vol. 2518, pp. 453–464. Springer, Berlin (2002)zbMATHGoogle Scholar
 4.Björklund, A., Husfeldt, T., Kaski, P., Koivisto, M.: Narrow sieves for parameterized paths and packings. arXiv:1007.1161 (2010)
 5.Björklund, A., Kaski, P., Kowalik, L.: Probably optimal graph motifs. In: 30th International Symposium on Theoretical Aspects of Computer Science (STACS 2013), LIPIcs, vol. 20, pp. 20–31, Schloss DagstuhlLeibnizZentrum fuer Informatik, Dagstuhl, Germany (2013)Google Scholar
 6.Björklund, A., Kaski, P., Kowalik, Ł.: Fast witness extraction using a Decision Oracle. In: Schulz, A.S., Wagner, D. (eds.) Algorithms–ESA 2014. ESA 2014. Lecture Notes in Computer Science, vol. 8737. Springer, Berlin, Heidelberg (2014)Google Scholar
 7.Björklund, A., Kaski, P., Kowalik, L., Lauri, J.: Engineering motif search for large graphs. In: Proceedings of the Seventeenth Workshop on Algorithm Engineering and Experiments (ALENEX 2015), SIAM, 2015, pp. 104–118 (2015)Google Scholar
 8.Creignou, N., Ktari, R., Meier, A., Müller, J.S., Olive, F., Vollmer, H.: Parameterized enumeration for modification problems. In: Dediu, A.H., Formenti, E., MartínVide, C., Truthe, B. (eds.) Language and Automata Theory and Applications, LATA 2015. Lecture Notes in Computer Science, vol. 8977. Springer, Cham (2015)Google Scholar
 9.Creignou, N., Meier, A., Müller, J.S., Schmidt, J., Vollmer, H.: Paradigms for parameterized enumeration. In: Chatterjee, K., Sgall, J. (eds.) Mathematical Foundations of Computer Science 2013, MFCS 2013. Lecture Notes in Computer Science, vol. 8087. Springer, Berlin, Heidelberg (2013)Google Scholar
 10.Curticapean, R.: Counting Matchings of Size k is #W[1]hard. In: Fomin, F.V., Freivalds, R., Kwiatkowska, M., Peleg, D. (eds.) Automata, Languages, and Programming. ICALP 2013. Lecture Notes in Computer Science, vol. 7965. Springer, Berlin, Heidelberg (2013)Google Scholar
 11.Curticapean, R., Marx, D.: Complexity of counting subgraphs: only the boundedness of the vertexcover number counts. In: 55th Annual IEEE Symposium on Foundations of Computer Science (FOCS 2014)Google Scholar
 12.Downey, R.G., Fellows, M.R.: Fundamentals of Parameterized Complexity. Springer, London (2013)CrossRefzbMATHGoogle Scholar
 13.Fernau, H.: On parameterized enumeration. In: Ibarra, O.H., Zhang, L. (eds.) Computing and combinatorics. COCOON 2002. Lecture Notes in Computer Science, vol. 2387. Springer, Berlin, Heidelberg (2002)Google Scholar
 14.Flum, J., Grohe, M.: The parameterized complexity of counting problems. SIAM J. Comput. 33(4), 892–922 (2004)MathSciNetCrossRefzbMATHGoogle Scholar
 15.Flum, J., Grohe, M.: Parameterized Complexity Theory. Springer, Berlin (2006)zbMATHGoogle Scholar
 16.Gelbord, B.: Graphical techniques in intrusion detection systems. In: Proceedings of 15th International Conference on Information Networking, pp. 253–258 (2001)Google Scholar
 17.Jerrum, M., Meeks, K.: The parameterised complexity of counting even and odd induced subgraphs. Combinatorica (2016). https://doi.org/10.1007/s0049301633385
 18.Jerrum, M., Meeks, K.: The parameterised complexity of counting connected subgraphs and graph motifs. J. Comput. Syst. Sci. 81(4), 702–716 (2015)MathSciNetCrossRefzbMATHGoogle Scholar
 19.Jerrum, M., Meeks, K.: Some hard families of parameterised counting problems. ACM Trans. Comput. Theory 7(3), 11 (2015)CrossRefzbMATHGoogle Scholar
 20.Khuller, S., Vazirani, V.V.: Planar graph coloring is not selfreducible, assuming P \(\ne \) NP. Theor. Comput. Sci. 88(1), 183–189 (1991)MathSciNetCrossRefzbMATHGoogle Scholar
 21.Lawler, E.L.: A procedure for computing the \(k\) best solutions to discrete optimization problems and its application to the shortest path problem. Manag. Sci. 18(7), 401–405 (1972)MathSciNetCrossRefzbMATHGoogle Scholar
 22.Meeks, K.: The challenges of unbounded treewidth in parameterised subgraph counting problems. Discrete Appl. Math. 198, 170–194 (2016)MathSciNetCrossRefzbMATHGoogle Scholar
 23.Milo, R., ShenOrr, S., Itzkovitz, S., Kashtan, N., Chklovskii, D., Alon, U.: Network motifs: simple building blocks of complex networks. Science 298(5594), 824–827 (2002)CrossRefGoogle Scholar
 24.Naor, M., Schulman, L.J., Srinivasan, A.: Splitters and nearoptimal derandomization. In: Proceedings of IEEE 36th Annual Foundations of Computer Science (FOCS 1995), Milwaukee, WI, 1995, pp. 182191. https://doi.org/10.1109/SFCS.1995.492475
 25.Sekar, V., Xie, Y., Maltz, D.A., Reiter, M.K., Zhang, H.: Toward a framework for internet forensic analysis. In: Third Workshop on Hot Topics in Networking (HotNetsIII) (2004)Google Scholar
 26.StanifordChen, S., Cheung, S., Crawford, R., Dilger, M., Frank, J., Hoagland, J., Levitt, K., Wee, C., Yip, R., Zerkle, D.: GrIDS—a graph based intrusion detection system for large networks. In: Proceedings of the 19th National Information Systems Security Conference, pp. 361–370 (1996)Google Scholar
 27.Schnorr, C.P.: Optimal algorithms for selfreducible problems. In: Proceedings of the 3rd ICALP, Edinburgh University Press, pp. 322 – 337 (1976)Google Scholar
Copyright information
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.