Small sample spaces with almost independent random variables are applied to design efficient sequential deterministic algorithms for two problems. The first algorithm, motivated by the attempt to design efficient algorithms for the All Pairs Shortest Path problem using fast matrix multiplication, solves the problem of computing witnesses for the Boolean product of two matrices. That is, if A and B are two n by n matrices, and C = AB is their Boolean product, the algorithm finds for every entry Cij = 1 a witness: an index k so that Aik = Bkj = 1. Its running time exceeds that of computing the product of two n by n matrices with small integer entries by a polylogarithmic factor. The second algorithm is a nearly linear time deterministic procedure for constructing a perfect hash function for a given n-subset of {1,..., m}.

... We show lower bounds that come close to our upper bounds (for a large range of n and ffl): Schemes that answer queries with just one bitprobe and error probability ffl must use \Omega ( nffl log(1=ffl) log m) bits of storage; if the error is restricted to queries not in S, then the scheme must use \Omega ( n2ffl2 log(n=ffl) log m) bits of storage. We also

We consider dictionaries of size n over the finite universe U = and introduce a new technique for their implementation: error correcting codes. The use of such codes makes it possible to replace the use of strong forms of hashing, such as universal hashing, with much weaker forms, such as clustering. We use

. We investigate the problem of storing a subset of the elements of a boundeduniverse so that searches canbe performed in constant time and the space used is within a constant factor of the minimum required. Initially we focus on the static version of this problem and conclude with an enhancement that permits insertions and deletions. 1 Introduction Given a universal set M = f0; : : : ; M \Gamma 1g and any subset N = fe 1 ; : : : ; e N g the membership problem is to determine whether given query element in M is an element of N . There are two standard approaches to solve this problem: to list all elements of N (e.g. in a hash table) or to list all the answers (e.g. a bit map of size M ). When N is small the former approach comes close to the information theoretic lower bound on the number of bits needed to represent an arbitrary subset of the given size (i.e. a function of both N and M , l lg \Gamma M N \Delta m ). Similarly, when N is large (say ffM ) the later approach is near...

We consider the problem of storing an n element subset S of a universe of size m, so that membership queries (is x 2 S?) can be answered efficiently. The model of computation is a random access machine with the standard instruction set (direct and indirect adressing, conditional branching, addition, subtraction, and multiplication). We show that if s memory registers are used to store S, where n s m=n , then query time \Omega\Gammame/ n) is necessary in the worst case. That is, under these conditions, the solution consisting of storing S as a sorted table and doing binary search is optimal. The condition s m=n is essentially optimal; we show that if n + m=n o(1) registers may be used, query time o(log n) is possible.

A minimal perfect hash function maps a set S of n keys into the set { 0, 1,..., n − 1} bijectively. Classical results state that minimal perfect hashing is possible in constant time using a structure occupying space close to the lower bound of log e bits per element. Here we consider the problem of monotone minimal perfect hashing, in which the bijection is required to preserve the lexicographical ordering of the keys. A monotone minimal perfect hash function can be seen as a very weak form of index that provides ranking just on the set S (and answers randomly outside of S). Our goal is to minimise the description size of the hash function: we show that, for a set S of n elements out of a universe of 2 w elements, O(n log log w) bits are sufficient to hash monotonically with evaluation time O(log w). Alternatively, we can get space O(n log w) bits with O(1) query time. Both of these data structures improve a straightforward construction with O(n log w) space and O(log w) query time. As a consequence, it is possible to search a sorted table with O(1) accesses to the table (using additional O(n log log w) bits). Our results are based on a structure (of independent interest) that represents a trie in a very compact way, but admits errors. As a further application of the same structure, we show how to compute the predecessor (in the sorted order of S) of an arbitrary element, using O(1) accesses in expectation and an index of O(n log w) bits, improving the trivial result of O(nw) bits. This implies an efficient index for searching a blocked memory.

A static dictionary is a data structure for storing a subset S of a finite universe U so that membership queries can be answered efficiently. We explore space efficient structures to also find the rank of an element if found. We first give a representation of a static dictionary that takes n lg m + O(lg lg m) bits of space and supports membership and rank (of an element present in S) queries in constant time, where n = jSj and m = jU j. Using our structure we also give a representation of a m-ary cardinal tree with n nodes using ndlg me + 2n + o(n) bits of space that supports the tree navigational operations in O(1) time, when m is o(2 ). For arbitrary m, we give a structure that takes the same space and supports all the navigational operations, except finding the child labeled i (for any i), in O(1) time. Finding the child labeled i in this structure takes O(lg lg lg m) time.

Abstract. We consider the issues of implicitness and cache-obliviousness in the classical dictionary problem for n distinct keys over an unbounded and ordered universe. One finding in this paper is that of closing the longstanding open problem about the existence of an optimal implicit dictionary over an unbounded universe. Another finding is motivated by the antithetic features of implicit and cache-oblivious models in data structures. We show how to blend their best qualities achieving O(log n) time and O(log B n) block transfers for searching and for amortized updating, while using just n memory cells like sorted arrays and heaps. As a result, we avoid space wasting and provide fast data access at any level of the memory hierarchy. 1

Abstract. We introduce an innovative decomposition technique which reduces a multi–dimensional searching problem to a sequence of one–dimensional problems, each one easily manageable in optimal time×space complexity using traditional searching strategies. The reduction has no additional storage requirement and the time complexity to reconstruct the result of the original multi–dimensional query is linear in the dimension. More precisely, we show how to preprocess a set of S ⊆ IN d of multi–dimensional objects into a data structure requiring O(m log n) space, where m = |S | and n is the maximum number of different values for each coordinate. The obtained data structure is implicit, i.e. does not use pointers, and is able to answer the exact match query in 7(d − 1) steps. Additionally, the model of computation required for querying the data structure is very simple; the only arithmetic operation needed is the addition and no shift operation is used. The technique introduced, overcoming the multi–dimensional bottleneck, can be also applied to non traditional models of computation as external memory, distributed, and hierarchical environments. Additionally, we will show how the proposed technique permits the effective realizability of the well known perfect hashing techniques on real data. The algorithms for building the data structure are easy to implement and run in polynomial time. 1

Abstract not found