Showing papers on "Crossover published in 1987"

A study of permutation crossover operators on the traveling salesman problem

Abstract: The application of Genetic Algorithms to problems which are not amenable to bit string representation and traditional crossover has been a growing area of interest. One approach has been to represent solutions by permutations of a list, and "permutation crossover" operators have been introduced to preserve legality of offspring. Three permutation crossovers are analyzed to characterize how they sample the o-schema space, and hence what type of problems they may be applicable to. Experiments performed on the Traveling Salesman Problem go some way to support the theoretical analysis.

Placement by Simulated Annealing on a Multiprocessor

Abstract: Physical design tools based on simulated annealing algorithms have been shown to produce results of extremely high quality, but typically at a very high cost in execution time. This paper selects a representative annealing application--standard cell placement--and develops multiprocessor-based annealing algorithms for placement. A taxonomy of possible multiprocessor decompositions of annealing algorithms is presented which divides decomposition schemes into two broad classes: those which divide individual moves into subtasks and distribute them across cooperating processors, and those which perform complete moves in parallel. It is shown that the choice of multiprocessor annealing strategy is influenced by temperature; in particular, the paper introduces the idea of adaptive strategies that dynamically change the parallel decomposition scheme to achieve maximum speedup as the annealing task progresses through each temperature regime. Implementations of three parallel placement strategies are described for an experimental shared-memory multiprocessor. Practical speedups are achieved over a serial version of the algorithm, and it is shown that an adaptive strategy which switches between two parallel decompositions at the optimal temperature yields speedup significantly better than any single strategy approach. Models are developed to account for the observed performance, and to predict the crossover points for switching strategies.

Using the genetic algorithm to generate LISP source code to solve the prisoner's dilemma

Abstract: The genetic algorithm is adapted to manipulate Lisp S-expressions. The traditional genetic operators of crossover, inversion, and mutation are modified for the Lisp domain. The process is tested using the Prisoner's Dilemma. The genetic algorithm produces solutions to the Prisoner's Dilemma as Lisp S-expressions and these results are compared to other published solutions.

Optimal crossover designs for the comparison of two treatments in the presence of carryover effects and autocorrelated errors

Abstract: SUMMARY A model for responses in a crossover experiment that includes direct and first-order carryover treatment effects, together with an autocorrelated error term, is considered when there are two treatments. A method for generating designs that minimize the variance of the estimated treatment effects is proposed and explicit results are given for the three and four period cases. Efficiencies of some practical designs are calculated.

Computer-aided pipeline operation using genetic algorithms and rule learning. PART I: Genetic algorithms in pipeline optimization

Abstract: In this two-paper series, techniques connected with artificial intelligence and genetics are applied to achieve computer-based control of gas pipeline systems. In this, the first paper, genetic algorithms are developed and applied to the solution of two classical pipeline optimization problems, the steady serial line problem, and the single transient line problem. Simply stated, genetic algorithms are canonical search procedures based on the mechanics of natural genetics. They combine a Darwinian survival of the fittest with a structured, yet randomized, information exchange between artificial chromosomes (strings). Despite their reliance on stochastic processes, genetic algorithms are no simple random walk; they carefully and efficiently exploit historic information to guide future trials. In the two pipeline problems, a simple three-operator genetic algorithm consisting of reproduction, crossover, and mutation finds near-optimal performance quickly. In, the steady serial problem, near-optimal performance is found after searching less than 1100 of 1.1(1012) alternatives. Similarly, efficient performance is demonstrated in the transient problem. Genetic algorithms are ready for application to more complex engineering optimization problems. They also can serve as a searning mechanism in a larger rule learning procedure. This application is discussed in the sequal.

Crossover counts and likelihood in multipoint linkage analysis.

Abstract: For large numbers of genetic loci, jointly tested to determine their order along a chromosome, likelihood methods become unfeasible owing to the very large numbers of discrete alternative hypotheses (locus orderings) whose likelihoods must be separately evaluated. A method to order loci according to the criterion of minimizing the obligatory crossover count is therefore proposed. A branch-and-bound algorithm implementing this proposal has been programmed; the properties of this algorithm are investigated. The statistical properties of the proposed method are also considered. It is shown to be consistent under wide conditions, including arbitrary locus spacings, variable amounts of information per locus, and some patterns of interference. The relationship between the minimum crossover order and the maximum likelihood order is discussed. For fully informative gametes, and tight linkage, there is a virtual equivalence of the two criteria. For looser linkage, there remains a close relationship.

Multidriver loudspeaker apparatus with improved crossover filter circuits

Abstract: A loudspeaker system includes at least two loudspeaker drivers, together with an electrical crossover network having filter circuits with at least two separate mutually exclusive frequency passbands. The filter circuits comprising the crossover network each possess brick-wall amplitude responses, i.e., passbands with very high band-edge amplitude vs. frequency response slops, on the order of 100 dB/octave in the better embodiments. The high passband band-edge slopes, which are realized by the inclusion of transmission zeros in the separate crossover filter transfer functions, takes together with further appropriate crossover filter transfer function synthesis causes the separate loudspeaker drivers comprising the loudspeaker system to function independently of one another in their contribution to total system acoustic output. It is shown that the loudspeaker system permits an accurate approximation to the ideal delay function in acoustic space, while minimizing acoustic wave interference among drivers operating in adjacent frequency band, and also reducing overall system nonlinear distortion.

Dimensionality crossover in superconducting wire networks.

Abstract: We have fabricated wire networks in the form of triangular arrays of third-order Sierpinski gaskets and percolation arrays on square lattices. For the gaskets the superconducting phase boundary, ${\mathrm{T}}_{\mathrm{c}}$(H), exhibits a clear crossover between the homogeneous and inhomogeneous (fractal) regimes. This is formally equivalent to the phonon-fracton crossover. Surprisingly, the random, square-lattice percolation networks, which are expected to exhibit a similar crossover, seem to behave in a manner characteristic of two-dimensional systems, independent of length scale.

Role of the central dinucleotide at the crossover sites for the selection of quasi sites in DNA inversion mediated by the site-specific Cin recombinase of phage P1

Abstract: The crossover sites for Cin-mediated inversion consist of imperfect 12 bp inverted repeats with non-palindomic dinucleotides at the center of symmetry. Inversion is believed to occur in vivo between the homologous central 2 bp crossover sequences at the inversely repeated crossover sites through introduction of 2 bp staggered cuts and subsequent reciprocal strand exchanges. The site-specific Cin recombinase acts not only on the normal crossover sites but also, less efficiently, on quasi crossover sites which have some homology with the normal sites. We identified 15 new quasi sites including 4 sites within the cin structural gene. Homology at the 2 bp crossover sequences between recombining sites favors selection as quasi crossover sites. The Cin enzyme can occasionally mediate inversion between nonidentical crossover sequences and such recombinations often result in localized mutations including base pair substitutions and deletions within the 2 bp crossover sequences. These mutations are explained as the consequences of heteroduplex molecules formed between the staggered dinucleotides and either tubsequent resolution by DNA replication or subsequent mismatch repair. Occasional utilization of quasi crossover sites and localized mutagenesis at the crossover sequences in enzyme-mediated inversion processes would be one of the mechanisms contributing to genetic diversity.

A class of magnitude complementary loudspeaker crossovers

Abstract: Necessary and sufficient conditions are developed for a pair of transfer functions to have magnitude responses which sum to a constant. If, in addition, the transfer function pair is constrained to be all-pass complementary, the transfer function pair so obtained is well suited for loudspeaker crossover applications. Such crossover systems are characterized by a pair of transfer functions which exhibit the same phase angle at all frequencies. This property has twofold implications: i) in active crossover biamplified audio systems, whereby the low frequencies and high frequencies are reproduced using separate power amplifiers, in-phase crossover transfer functions require less power in subsequent amplifier stages than other designs to achieve a given acoustic sound pressure level; and ii) the summed acoustic magnitude response is least sensitive to noncoincidental mounting of the low- and high-frequency transducers when the crossover transfer functions exhibit the same phase angle at all frequencies. The class of transfer functions realizable is quite wide, and includes squared versions of Butterworth, Chebyshev, and elliptic transfer functions of all orders.

Crossover in the Anderson transition: Acoustic localization with a flow

Abstract: Acoustic waves propagating in a sufficiently disordered medium are localized because of multiple elastic scattering. The presence of a uniform flow in the medium breaks time-reversal invariance and changes the mechanism of localization. A field theory incorporating the crossover induced by a flow field is described and the crossover behavior of the localization length is calculated. An experimental realization of this phenomenon is proposed using third sound in a superfluid film on a rough substrate with a uniform superfluid flow.

Migraine Trials: Crossover or Parallel Group?

Abstract: This paper evaluates the relative merits of the crossover design and the parallel group design for the evaluation of migraine therapy, focussing on prophylaxis. Three issues are highlighted which lead

Patient Numbers Needed in Prophylactic Migraine Trials

Abstract: Based on a multicenter double-blind crossover prophylatic migraine trial (n = 96), the relative effectiveness of the crossover design versus group comparison was evaluated. If a ‘time effect’ is taken into account in the statistical model, the crossover design was estimated to be 10–11 times more effective than the group comparison. With more traditional statistical methods using paired and unpaired t tests, the crossover design was estimated to be 4 times more powerful than the group comparison. Based on the t test, sample size tables for numbers of patients needed for detecting effect differences from 10 to 50% with powers from 0.6 to 0.9 in a crossover and a group comparison are presented. These calculations are based on only one trial, and it is recommended that similar estimations of sample variability should be done in other previous trials.

On variance estimation in crossover designs.

Abstract: It is very unlikely that the errors in crossover experiments with more than two treatments are uncorrelated. The assumption of uncorrelated errors generally leads to underestimation of the variances of estimates. This paper determines bounds for the degree of underestimation for arbitrary covariance matrices, provided the experiment is planned according to a design that has the same numbers of periods and treatments and is balanced for carryover effects.

An information theoretic interface for a stochastic model management system

Abstract: Groups using group decision support systems GDSS for addressing organizational problems is an evolutionary process. An analytical model incorporating evolutionary processes exists, capturing this adaptation in the group decision-making . process. This model is based on the genetic algorithm GA and can be used to estimate GA parameter values from experimental data. This research effort examines possible relationships between the GA crossover and mutation parameters and the group context variables of leadership. Both the presence of and the activity level of group leaders are considered. Particular attention is paid to model implementation for a specific instance of GDSS use. The results of this effort are generally encouraging, hinting at the need to conduct further research in this area. q 2000 Elsevier Science B.V. All rights reserved.

Single-Row Routing with Crossover Bound

Abstract: Previous studies of the single-row routing problem have been restricted to the minimization of the total number of horizontal tracks needed for the realization of a given set of nets. Therefore, it has been assumed that enough space exists between adjacent nodes to allow for the wiring. Due to this assumption, realizations obtained with previously proposed algorithms may require a large number of vertical tracks between adjacent nodes. In this paper, we study the single-row routing problem when the number of vertical tracks available between adjacent nodes is bounded by a positive integer called the crossover bound. We give some results concerning crossovers and prove that, for any given positive integer K, an instance can be constructed such that the vertical track requirement between adjacent nodes cannot be less than K. We develop a fast algorithm for the case when the number of horizontal tracks available as well as the number of vertical tracks available between adjacent nodes have been preset. We compare the performance of our algorithm to the performance of an algorithm (proposed in [6]) which is fast and does not consider the vertical track constraint. Our experiments show that, in all cases, the realizations found by our algorithm have the same street capacities as those obtained by the algorithm proposed in [6]. However, unlike the realizations found with the algorithm proposed in [6], the ones found by our algorithm have smaller crossover bounds. The computing time of our algorithm is, in general, no worse than the computing time of the algorithm proposed in [6].

crossover of interfacial dynamics

Abstract: Some dynamical models in which there is a significant interplay between interfacial and bulk degrees of freedom are treated at the mean-field (or Van Hove) level from a coarse-grained microscopic viewpoint. Specifically, the near-equilibrium interfacial dynamics of two of the simplest models with conservation laws, models B and C, are studied with use of (as appropriate) variational techniques, perturbation theory, and (with certain additional simplifications) exact solutions. Use of these methods allows the dispersion relation for interfacial modes to be interpolated between the \"hydrodynamical\" and critical regimes. The crossover scaling behavior lends support to renormalizationgroup methods near d=1 which focus on the interfacial modes but nonetheless extract the bulk dynamical exponent as well as the crossover.

The Ising crossover between multi-layer and single-layer Sierpinski gaskets

Abstract: The authors propose a new fractal lattice-a multi-layer Sierpinski gasket with infinite ramification. The critical behaviour of an Ising model of this is investigated. In particular, the crossover behaviour is shown in a renormalisation group flow diagram.

Genetic algorithms and their applications : proceedings of the Second International Conference on Genetic Algorithms, July 28-31, 1987, at the Massachusetts Institute of Technology, Cambridge, MA

Abstract: ContentsGenetic Search Theory. D.E. Goldberg, P. Segrest, Finite Markov Chain Analysis of Genetic Algorithms. C.L. Bridges, D.E. Goldberg, An Analysis of Reproduction and Crossover in a Binary-Coded Genetic Algorithm. J.E. Baker, Reducing Bias and Inefficiency in the Selection Algorithm. T.H. Westerdale, Adaptive Search Operators I.Altruism in the Bucket Brigade. I. StadnykSchema Recombination in Pattern Recognition Problems. J.D. Schaffer, A. Morishima, An Adaptive Crossover Distribution Mechanism for Genetic Algorithms. D.E. Goldberg, J. Richardson, Representation Issues.Genetic Algorithms With Sharing for Multimodal Function Optimization. C.G. Shaefer, The ARGOT Strategy: Adaptive Representation Genetic Optimizer Technique. D.E. Goldberg, R.E. Smith, Nonstationary Function Optimization Using Genetic Algorithms With Dominance and Diploidy. H.J. Antonisse, K.S. Keller, Genetic Operators for High-Level Knowledge Representations. A.S. Bickel, R.W. Bickel, Keynote Address.Tree Structured Rules in Genetic Algorithms. J.H. Holland, Adaptive Search Operators II.Genetic Algorithms and Classifier Systems: Foundations and Future Directions. G.E. Liepins, M.R. Hilliard, M. Palmer, M. Morrow, Greedy Genetics. J.Y. Suh, D. Van Gucht, Incorporating Heuristic Information Into Genetic Search. D. Whitley, Using Reproductive Evaluation to Improve Genetic Search and Heuristic Discovery. D.J. Sirag, P.T. Weisser, Connectionism and Parallelism I.Toward a Unified Thermodynamic Genetic Operator. C.P. Dolan, M.G. Dyer, Toward the Evolution of Symbols. D.G. Oosthuizen, SUPERGRAN: A Connectionist Approach to Learning, Integrating, Genetic Algorithms and Graph Induction. G.G. Robertson, Parallel Implementation of Genetic Algorithms in a Classifier System. J.P. Cohoon, S.U. Hegde, W.N. Martin, D. Richards, Parallelism II.Punctuated Equilibria: A Parallel Genetic Algorithm. C.B. Pettey, M.R. Leuze, J.J. Grefenstette, A Parallel Genetic Algorithm. A.V. Sannier, II, E.D. Goodman, Genetic Learning Procedures in Distributed Environments. P. Jog, D. Van Gucht, Parallelisation of Probablistic Sequential Search Algorithms. R. Tanese, Credit Assignment and Learning.Parallel Genetic Algorithms for a Hypercube. R.L. Riolo, Bucket Brigade Performance: I. Long Sequences of Classifiers. R.L. Riolo, Bucket Brigade Performance: II. Default Hierarchies. J.J. Grefenstette, Multilevel Credit Assignment in a Genetic Learning System. K.A. De Jong, Applications I.On Using Genetic Algorithms to Search Program Spaces. D.P. Greene, S.F. Smith, A Genetic System for Learning Models of Consumer Choice. I.M. Oliver, D.J. Smith, J.R.C. Holland, A Study of Permutation Crossover Operators on the Traveling Salesman Problem. M.R. Hilliard, G.E. Liepins, M. Palmer, M. Morrow, J. Richardson, A Classifier Based System for Discovering Scheduling Heuristics. C. Fujiko, J. Dickinson, Applications II.Using the Genetic Algorithm to Generate LISP Source Code to Solve the Prisoner's Dilemma. V.V. Raghavan, B. Agarwal, Optimal Determination of User- Oriented Clusters: An Application for the Reproductive Plan. S.W. Wilson, The Genetic Algorithm and Biological Development. L. Davis, S. Coombs, Genetic Algorithms and Communication Link Speed Design: Theoretical Considerations. S. Coombs, L. Davis, Genetic Algorithms and Communication Link Speed Design: Constraints and Operators.

Crossover detecting method for stacked fuel cell

Abstract: PURPOSE:To make it possible to simply detect crossover even while a cell is operated by measuring the output voltage in two portions of a fuel cell and detecting a degree of crossover in adjacent fuel cells by the voltage difference between two portions. CONSTITUTION:When local current 8 produced by crossover flows, a current flow direction is forcibly curved. In up and down cells 11, 13 adjacent to a cell 12 in which crossover arises, a large amount of current flows in a portion different from that in normal condition. Since the voltage within a cell can not reach equilibrium because of large area, the voltage in an inlet differs from that in an outlet. Therefore, the output voltage in the inlet 2 and that in the outlet 3 of a fuel cell 1 are measured and the crossover in the adjacent cells is detected by the voltage difference measured. Thereby, the crossover can simply be detected while the fuel cell is operated.

Randomization and Additivity in the Two‐Period Crossover Clinicial Trial

Abstract: In certain areas of medical research, the two-period crossover design is a frequent choice for comparing treatments A and B in a randomized clinical trial. Earlier work by Grizzle and by Brown was based upon a parametric theory linear model. Recently, the present authors employed D. R. Cox's additive randomization models and, for the case of zero residual effect, found a discrepancy between it and the parametric model with respect to the precision of period effects. In the present note, this divergence is accounted for by allowing for the possibility of non-additivity through the use of a completely general randomization model. It is concluded that the structure of the crossover design is such that use of the parametric theory linear model is required if a single, consistent model is desired.

Cell renormalisation study of biased self-avoiding walks

Abstract: The authors present a cell renormalisation approach for the biased self-avoiding walks and show that a stiff-to-isotropic crossover exponent is exactly one in all dimensions and for all cell sizes. They also show, by use of renormalisation flow diagrams, a substantial difference between two and three dimensions in the crossover from stiff limit to isotropic limit as the length of walk N to infinity for fixed gauche weight p. In three dimensions, a crossover seems to occur first to the random walk limit and then to the self-avoiding walk limit, while, in two dimensions, it seems to occur directly to the self-avoiding walk limit in agreement with recent observations based on Monte Carlo simulations.

The ƒ(a) formalism applied to the crossover from conservative to dissipative scaling behaviour in period doubling systems

Abstract: We use the ƒ( a ) (spectral dimension) formalism to give a global description of the crossover from conservative to dissipative scaling behaviour in period doubling systems. As a generic model we consider the two-dimensional Henon map, for which the jacobian 0 ⩽ B ⩽ 1 measures the amount of dissipation, and we calculate the successive approximants ƒ k ( a , B ) to the generalized spectral dimension for small (i.e. experimentally feasible) values of k . The crossover properties from the extreme dissipative case ( B = 0) to the conservative case ( B = 1) are discussed.

Scaling Theories for Anomalous Dynamics on Fractals: Fractons

Abstract: Many systems in nature exhibit different geometrical structures at different length scales. Although they are homogeneous (on the average) for large length scales, they exhibit self-similarity on short length scales. In the latter case such systems may be modeled by fractal structures.1 On a fractal structure, all the physical properties behave as powers of the relevant length scale. This behavior crosses over to a homogeneous one, similar to that of usual bulk matter, at a crossover length, called ξ. Assuming that ξ is the only relevant length scale in the problem, the dependence on any other length L should arise via the ratio L/ξ. A typical intrinsic quantity χ will have the “fractal” behavior χ~Lx for L«ξ, and will crossover to a size independent behavior, χ~ξx, for L>>ξ.

Time domain noise-induced jitter: Theory and precise measurement

Abstract: A theoretical relationship between recording channel signal-to-noise ratio and the time distribution of signal crossover (bitshift distribution) is derived. A technique for measuring crossover distributions with 5 picosecond resolution is described.