simulation@uflorida.cis.ufl.edu (Moderator: Paul Fishwick) (01/09/89)
Volume: 7, Issue: 2, Mon Jan 9 09:12:57 EST 1989 +----------------+ | TODAY'S TOPICS | +----------------+ (1) Genetic Algorithms (2) Supercomputing/Numerical Aerodynamics Simulation Division (3) Looking for BNF Grammar for SPICE (4) Electronic Mail Address for SIMULA * Moderator: Paul Fishwick, Univ. of Florida * Send topical mail to: simulation@uflorida.cis.ufl.edu * Archives available via FTP to bikini.cis.ufl.edu, login as 'anonymous', use your last name as the password, change directory to pub/simdigest. ----------------------------------------------------------------------------- Date: Mon, 19 Dec 88 09:33:25 EST From: Paul Fishwick <fishwick@fish.cis.ufl.edu> To: simulation@ufl.edu [[FROM EDITOR: taken from news group 'bionet.molbio.evolution' -paf]] Newsgroups: bionet.molbio.evolution Subject: Simulating Evolution Date: 17 Dec 88 22:52:00 GMT References: <4806@phoenix.Princeton.EDU> <4029b61d.ffb5@bumper.engin.umich.edu> <7034@venera.isi.edu> Reply-To: offutt@caen.engin.umich.edu (daniel m offutt) Organization: caen In article <7034@venera.isi.edu> smoliar@vaxa.isi.edu.UUCP (Stephen Smoliar) writes: >The characterization of genetic algorithms may have reversed the cart and the >horse. Genetic algorithms may be said to have been inspired by chromosomal >behavior, but I think it would be an exaggeration to call them a simulation, >of any organic situation. Genetic algorithms are not just motivated by evolution, they are simulations of evolution including: 1. A population of linear or circular strands of DNA each modeled using a two-letter alphabet, 2. An environment modeled as a computable function which determines fitness of a chromosome, 3. Fitness-based reproduction of chromosomes (offspring), 4. Crossing over of pairs of chromosomes producing recombinants, 5. Point mutation, 6. Inversion (in some models), 7. Punctuation marking sites of possible crossing over (in some models), 8. Recessive and dominant alleles (in some models), 9. Speciation (in some models), 10. Semi-isolated subpopulations (in some models), 11. Mate-selection procedures (in some models), 12. A temporal ordering and causal connectedness among these events which correspends to the modeled events in nature: successive generations of creation, testing, recombination, reproduction, and finally deletion from the population. The world is too complex for any computer simulation to capture every aspect of evolution, or any other physical process. Consequently there are aspects of evolution not simulated (yet) in genetic algorithms, including: 1. Fertilization, 2. Spontaneous abortion, 3. Multiple chromosomes in one genotype, 4. Any of the idiosyncrasies of the generation of sex cells aside from crossing over. 5. Homologous pairs. 6. Any of the codon sequences found in DNA, aside from those coding for sites of possible crossing over. 7. Transcription. 8. Protein synthesis. 9. etc., etc. My point again is that the algorithm was copied from nature, not invented by, say, a clever operations researcher. (The formal theory of genetic algorithms is another matter. Holland (1975,1976) has proven a generalization of Fisher's (1930) classic result which is not restricted to alleles, but holds for all subsets of alleles within individual chromosomes. Genetic algorithms derive largely from this genetic search theory (GST) which falls within the field of mathematical genetics.) A "no-frills" genetic algorithm, with the basic allele-set ranking capabilities that make GA's effective function optimizers, can be implemented in one page of C code. The fundamental algorithm is simple enough that it is quite plausible that it was discovered (by nature) via a much weaker pure-random search process. The observed efficiency of genetic algorithms at locating ever-better local optima on a wide variety of complex performance surfaces, without becoming trapped on a local optimum anywhere, counts as evidence that if it were applied to optimizing performance measures defined by natural environments, it would be able to locate quite complex stable physical systems such as plants and animals. So perhaps, given the similarity of the algorithm to the natural evolutionary process, this is what has actually happened. ====================================================================== Dan Offutt offutt@caen.engin.umich.edu ------------------------------ To: ames!comp-simulation@ames.arc.nasa.gov Path: eos!eugene From: eos!eugene@eos.arc.nasa.gov (Eugene Miya) Newsgroups: comp.simulation Subject: Supercomputing: Experience Preferred But Not Essential (Announcement of Opportunity) Keywords: university/college graduating students Date: 5 Jan 89 07:14:31 GMT Sender: eos!eugene@eos.arc.nasa.gov Reply-To: eos!eugene@eos.arc.nasa.gov (Eugene Miya) Distribution: usa Organization: NASA Ames Research Center, Calif. Are you a savvy tiger? A Special Announcement of Opportunity (AO) Numerical Aerodynamic Simulation Systems Division NASA Ames Research Center The Numerical Aerodynamic Simulation (NAS) Systems Division is seeking new, recent, "fresh-out," graduates for challenging positions in large-scale computer systems research, development, and operational support at the NASA Ames Research Center located on Moffett Field, CA [southern-most tip of San Francisco Bay in the Santa Clara Valley]. The NAS Facility houses one of the world's largest and most advanced supercomputer systems, used by a nationwide user community for solving highly complex problems in fluid dynamics, aeronautics, space, and other scientific disciplines. The System includes a CRAY Y-MP*, a CRAY-2, a Thinking Machine's Connection Machine-2, dual Amdahl 5880's, four VAX 11/780s, over 30 Silicon Graphics Iris and SUN workstations interconnected with local area networks ranging in speed 10 to 800 Mhz. The user visible operating system is UNIX with TCP/IP-based networking extensions. [Everything!] Systems research, development, and computational service opportunities exist in the following areas: 1) Supercomputers 2) Mass Storage Systems 3) High Performance Scientific Workstations 4) High Performance Data Networks 5) Long Haul Communications 6) Mini-supercomputers Future developments will involve NeXT machines, Ardent Titan and Stellar graphical superworkstations, and other flavors of state of the art high technology. These positions require at least a Bachelor's degree [or more] in computer science, engineering, or one of the physical sciences. Knowledge of the UNIX operating system and the DoD Internet Protocols is highly desireable. This is an excellent entry opportunity for the young and young-at-heart. NASA (The National Aeronautics and Space Administration) is the Nation's civilian space Agency. The NASA Ames Research Center is an Equal Opportunity Employer. US Citizenship is required. Security clearance is not required. Send resumes to: Mr. Bruce Blaylock Chief, NAS Systems Development Branch M/S 258-5 NASA Ames Research Center Moffett Field, CA 94035 blaylock@prandtl.nas.nasa.gov or ames!prandtl!blaylock TeX and troff okay. *Trademarks: CRAY Y-MP, CRAY-2 are trademarks of Cray Research Inc. Connection Machine-2 is a trademark of Thinking Machines Corp. UNIX is a trademark of AT&T Bell Laboratories. VAX is a trademark of Digital Equipment Corporation. Iris is a trademark of Silicon Graphics Inc. SUN is a trademark of SUN Microsystems Inc. Send specific questions to me. eugene miya, NASA Ames Research Center, eugene@aurora.arc.nasa.gov resident cynic at the Rock of Ages Home for Retired Hackers: "Mailers?! HA!", "If my mail does not reach you, please accept my apology." {uunet,hplabs,ncar,decwrl,allegra,tektronix}!ames!aurora!eugene ------------------------------ To: comp-simulation@rutgers.edu Path: hplabs!hp-ses!hpcuhb!najmi From: hplabs!najmi@hpcuhb.hp.com (Farrukh Najmi) Newsgroups: comp.simulation Subject: Looking for BNF for SPICE. Date: 6 Jan 89 17:00:45 GMT Organization: Hewlett Packard, Cupertino I am looking for a BNF description of HSPICE or spice syntax. If any one has a yacc parser and lex scanner for any variation of spice I would love to have it. Thanks in advance. Farrukh Najmi. ------------------------------ Sender: "Leonard_S._Li.ESAE"@Xerox.COM Date: 6 Jan 89 11:30:56 PST (Friday) Subject: Electronic mail address for Simula a.s.? From: LLi.ESAE@Xerox.COM To: Simulation@ufl.EDU Cc: LLi.ESAE@Xerox.COM We are interested in the capabilities and the cost of Simula. The 10/88 issue of "Simulation" lists the address of the organization supporting Simula as: Simula a.s. Postbox 4403 Torshov N-0402 Oslo 4 Norway (472) 156710 Does anyone know if they have an electronic mail address? Please also tell us if they have an agent/dealer in the U.S. Thanks. Leonard Li Xerox El Segundo, CA, U.S.A. ------------------------------ +--------------------------+ | END OF SIMULATION DIGEST | +--------------------------+