arXiv subjects

J. Boutet de Monvel

Publications and source records attributed to J. Boutet de Monvel.

3 recordsLinked to original sources

arXiv2000

Mean-Field Approximations to the Longest Common Subsequence Problem

The Longest Common Subsequence (LCS) problem is a fundamental problem of sequence comparison. A natural approximation to this problem is a model in which every pairs of letters of two ``sequences'' are matched independently of the other pairs with probability 1/S, $S$ representing the size of the alphabet. This model is analogous to a mean field version of the LCS problem, which can be solved with a cavity approach (Eur. Phys. J. B 7-2(1999),pp. 293-308). We refine here this approximation by incorporating in a systematic way correlations among the matches in the cavity calculation. We obtain a series of closer and closer approximations to the LCS problem, which we quantify in the large $S$ limit, both with a perturbative approach and by Monte-Carlo simulations. We find that, as it happens in the expansion around mean-field for other disordered systems, the corrections to our approximations depend upon long-ranged correlation effects which render the large $S$ expansion non-perturbative.

cond-mat.dis-nn↗

arXiv1998

Extensive Simulations for Longest Common Subsequences: Finite Size Scaling, a Cavity Solution, and Configuration Space properties

The Longest Common Subsequence (LCS) Problem asks for the longest sequence of (non-contiguous) matches between two given strings of characters. Using extensive Monte Carlo simulations, we find a finite size scaling law of the form E(L)/N =C + A/(N^1/2 ln N)+... for the mean LCS length of two random strings of size N over S letters. We provide precise estimates of C for S between 2 and 15. We consider also a related Bernoulli Matching model where the different entries of an N times M array are independently occupied with probability 1/S. In that case we find the expression of the limit of L(N,M)/N as N grows to infinity, as a function of r=M/N. This expression provides a very good approximation for the Random String model, which gets more and more accurate as S increases. The question of the ``universality class'' of the LCS problem is also considered. For the Bernoulli Matching model we find very good agreement with recent scaling predictions of Hwa and Lassig for Needleman-Wunsch sequence alignment. We find however that the variance of the LCS length has a different scaling different in the Random String model, suggesting that long-ranged correlations among the matches are relevant in this model. We finally study the ``ground state'' properties of this problem. We find in particular that the number of solutions typically grows exponentially with N, i.e. this system has a residual entropy at T=0. Also the overlap between two LCSs chosen at random is found to be self averaging and to aproach a definite value q(S)<1 as N grows.

cond-mat.dis-nn↗

arXiv1998

The random link approximation for the Euclidean traveling salesman problem

The traveling salesman problem (TSP) consists of finding the length of the shortest closed tour visiting N ``cities''. We consider the Euclidean TSP where the cities are distributed randomly and independently in a d-dimensional unit hypercube. Working with periodic boundary conditions and inspired by a remarkable universality in the kth nearest neighbor distribution, we find for the average optimum tour length = beta_E(d) N^{1-1/d} [1+O(1/N)] with beta_E(2) = 0.7120 +- 0.0002 and beta_E(3) = 0.6979 +- 0.0002. We then derive analytical predictions for these quantities using the random link approximation, where the lengths between cities are taken as independent random variables. From the ``cavity'' equations developed by Krauth, Mezard and Parisi, we calculate the associated random link values beta_RL(d). For d=1,2,3, numerical results show that the random link approximation is a good one, with a discrepancy of less than 2.1% between beta_E(d) and beta_RL(d). For large d, we argue that the approximation is exact up to O(1/d^2) and give a conjecture for beta_E(d), in terms of a power series in 1/d, specifying both leading and subleading coefficients.

cond-mat↗