| Related sites for http://www.mth.uea.ac.uk/~h090/primeEDS.html |
| CellServ Hands-on laboratory training kits for High School and College Students for unique laboratory experiences and instruction in biology, genetics, cancer, and biotechnology. | | Rockgrout Animated and interactive programs explaining the design and construction of cement grouting. Includes an instructive game in which an imaginary hole is 'grouted'. | | Life_Studies A beginner's guide to gold panning. Learn where the gold is, submersion, panning, separation and retrieval procedures. | | Bruker__Report_/_SpinReport_—_Online Several issues of the Bruker Report. There are articles about shimming, EPR and NQR. Technical information including "how to" articles. | | Bustamante_Group Laboratory of Carlos Bustamante at UC-Berkeley. Research focuses on the dynamics, structure, and kinetics of molecular motors and nucleo-protein assemblies. | | LMS_Technologies Specializes in offering filters and advice for air, paint spraying and liquid environments. Includes list of products and services. In Bloomington, Minnesota. | | Planning_Consultants_Online_Directory Directory of planning consultancies based within the United Kingdom. | | The_Human-Languages_Page The Human-Languages Page is a comprehensive catalog of language-related Internet resources. The over 1900 links in the HLP database have been hand-reviewed to bring the best language links the Web has | | Insects_of_Alberta Photo gallery of insects found in Alberta, Canada by R. Bercha. | | Acadia_University The Centre for Estuarine Research focuses research attention on the estuaries and nearshore coastal waters of Eastern Canada, with emphasis on the estuarine systems of the Bay of Fundy and the hydrogr | | Brown_Journal_of_World_Affairs Semi-annual international affairs journal published at the Watson Institute of International Relations, Brown University. Articles since 1996 available as PDF. | | Arizona_State_University Department of Mechanical and Aerospace Engineering | | Cascade_Technologies Developer and manufacturer of quantum cascade lasers for gas sensing and other applications. Real time measurement, parts per trillion, room temperature operation. | | Case_in_Old_English Robert Stevick's study of case and other grammatical elements of Old English. | | Chicago_Landmarks__Daniel_Burnham Biography and information about his projects in the city. | | NanoWorld_AG Swiss company that provides silicon and silicon nitride cantilevers and probes for SPM and AFM | | Swiss_Federal_Institutes_of_Technology_-_Physics_Department Contains links to the various departments and an organization chart. Contains limited English. | | 8th_International_Conference_on_New_Trends_in_Immunosuppression_and_Immunotherapy February 14-17, 2008, Berlin, Germany | | Relativistic_Astrophysics_Department_at_Sternberg_Astronomical_Institute Relativity group at Moscow State University. Pages give information on staff members, the group's history, and seminars. Research includes the physics of neutron stars, accretion disks, and cosmology. | | Palm_Haven Photographs and information on palm trees, cycads, bromeliads and tropical plants and advice on cultivation. |
|
Prof Graham R Everest, MTH, UEA Prime Values of Elliptic Divisibility Sequences If u[n] denotes the n-th term of an EDS then u[n] grows very quickly. In fact log|u[n]| ~ cn2.Chudnovsky and Chudnovsky considered some EDSs specified by writing down their first five terms and examined these sequences for prime terms. Since u[n] is a divisibility sequence, which grows so quickly, it is sufficient to examine prime occurence of terms u[n] with n prime. The sequences considered by the Chudnovskys are specified below by giving the first 5 terms, as well as the constant c and the occurrence of prime values for prime n up to 100. 0,1,1,1,-20.05605,7,11,13,23,61,710,1,1,1,60.11075,7,13,23,43,470,1,2,1,4 0.12625,7,710,1,1,2,70.131111,17,730,1,1,1,-90.13837,47,790,1,1,1,100.14327,13,41,610,1,1,4,10.173071,790,1,1,4,30.1737 5,7,13,53,710,1,1,5,2 0.20107,43Some of the primes in this table are very large. For example, the term u[79] in the third sequence from the end is a prime with 469 decimal digits. It might look as though we should be able to keep computing terms and find larger and larger primes. But if you run the sequences out to n=500 you find no new primes. In a recent paper, there is a heuristic explanation of why these sequences should stop producing primes beyond a certain point. This also explains why there shouldbe a uniform bound on the number of primes. A proof of finiteness under a hypothesis on 2-torsion is obtainablehere.Other ModelsSuppose we ask the same question for an elliptic curve in homogeneous form,x3 + y3 = c,for a non-zero rational c. Suppose P is a non-torsion rational point. Write x(nP)=An/Bn for integral An and Bn. The paper just referred to gives a proof that only finitely many terms Bn are prime.Zsigmondy's TheoremFor the Mersenne sequence Mn=2n-1, every term has a primitive divisor after n=6. A primitive divisor of Mn is a prime divisor which does not divide Mm for any smaller m.In a recent paperIgor Shparlinsky and I prove Zsigmondy's Theorem for EDSs.Primes From Rational PointsIt does look, in some cases, as though rational points on elliptic curves can produce primes in abundance if the Mordell-Weil rank of the curve is greater than 1. Let E denote such a curve, with two independent rational points P and Q. Let M and N denote integers and consider the bi-sequence MP+NQ, the denominator of the x-coordinate is always the square of an integer, let s(M,N) denote the square root of the denominator. We appear to have more joy looking for prime values of this bi-sequence. Simple heuristics (see below) suggest that the number of prime values of the sequence s(M,N) with |M| and |N| bounded by X is approximately clogX.The constant c should depend upon the elliptic regulator. The paper just referred to gives examples where only finitely many primes occur. For curves in homogeneous form, only finitely many primes appear with no extra hypotheses on the torsion. Nailing precise criteria for finiteness is an open problem. Cremona has a list of curves and points for conductors up to 6000. From these, the first few with rank 2 can be siphoned off to a separate list . Using this list, Peter Rogers is now collecting some data on this problem. Whether the heuristic argument is accurate or not, this method certainly produces some large primes.Examplese = [0,1,0,-25,39], P = [-5,8], Q = [-1,8]: s(118,31) is a prime with 2705 decimal digits. This curve appears as number 26 in the list . e = [0,-1,1,-9,9], P = [1,0], Q = [7,15]: s(111,47) is a prime with 2541 decimal digits. This curve is number 34 on the list. Heuristics for Prime OccurrenceThere is an interesting history of heuristics associated to prime occurrence in integer sequences. Chris Caldwell gives a brief account of heuristics for prime occurrence in the Mersenne Sequence in his interesting Prime Pages. The basic idea is to reckon that the Prime Number Theorem gives the probability 1/logN that the integer N is prime. So if you have an increasing sequence a[n] of positive integers then you would reckon that the number of terms of a[n] with n < X which are prime should be about S 1/loga[n],summed over n < X. Hardy and Wright pointed out that this kind of argument suggests there should only be finitely many Fermat Primes. In general, this basic argument needs some refinement to make it work. The numbers a[n] could all be even for example in which case the argument would be nonsense! In specific cases, such as Mersenne, refinements can be made to work which fit the data. In that case, Merten's Theorem is used to predict approximately vlogXMersenne Primes 2n-1 with n < X, where the constant v is given explicitly by the formula v = exp(g)/log2,where g is the Euler-Mascheroni constant. The paper gives a similar refinement for EDSs, using a combination of Merten's Theorem and Hasse's Theorem. This page maintained by: g.everest@mth.uea.ac.uk |
|
