Artificial Life Experiments Show How Complex Functions Can Evolve

NSF ^ | May 8, 2003 | Staff

[Nebullis]

Artificial Life Experiments Show How Complex Functions Can Evolve

Arlington, Va.—If the evolution of complex organisms were a road trip, then the simple country drives are what get you there. And sometimes even potholes along the way are important.

An interdisciplinary team of scientists at Michigan State University and the California Institute of Technology, with the help of powerful computers, has used a kind of artificial life, or ALife, to create a road map detailing the evolution of complex organisms, an old problem in biology.

In an article in the May 8 issue of the international journal Nature, Richard Lenski, Charles Ofria, Robert Pennock, and Christoph Adami report that the path to complex organisms is paved with a long series of simple functions, each unremarkable if viewed in isolation. "This project addresses a fundamental criticism of the theory of evolution, how complex functions arise from mutation and natural selection," said Sam Scheiner, program director in the division of environmental biology at the National Science Foundation (NSF), which funded the research through its Biocomplexity in the Environment initiative. "These simulations will help direct research on living systems and will provide understanding of the origins of biocomplexity."

Some mutations that cause damage in the short term ultimately become a positive force in the genetic pedigree of a complex organism. "The little things, they definitely count," said Lenski of Michigan State, the paper's lead author. "Our work allowed us to see how the most complex functions are built up from simpler and simpler functions. We also saw that some mutations looked like bad events when they happened, but turned out to be really important for the evolution of the population over a long period of time."

In the key phrase, "a long period of time," lies the magic of ALife. Lenski teamed up with Adami, a scientist at Caltech's Jet Propulsion Laboratory and Ofria, a Michigan State computer scientist, to further explore ALife.

Pennock, a Michigan State philosopher, joined the team to study an artificial world inside a computer, a world in which computer programs take the place of living organisms. These computer programs go forth and multiply, they mutate and they adapt by natural selection.

The program, called Avida, is an artificial petri dish in which organisms not only reproduce, but also perform mathematical calculations to obtain rewards. Their reward is more computer time that they can use for making copies of themselves. Avida randomly adds mutations to the copies, thus spurring natural selection and evolution. The research team watched how these "bugs" adapted and evolved in different environments inside their artificial world.

Avida is the biologist's race car - a really souped up one. To watch the evolution of most living organisms would require thousands of years – without blinking. The digital bugs evolve at lightening speed, and they leave tracks for scientists to study.

"The cool thing is that we can trace the line of descent," Lenski said. "Out of a big population of organisms you can work back to see the pivotal mutations that really mattered during the evolutionary history of the population. The human mind can't sort through so much data, but we developed a tool to find these pivotal events."

There are no missing links with this technology.

Evolutionary theory sometimes struggles to explain the most complex features of organisms. Lenski uses the human eye as an example. It's obviously used for seeing, and it has all sorts of parts - like a lens that can be focused at different distances - that make it well suited for that use. But how did something so complicated as the eye come to be?

Since Charles Darwin, biologists have concluded that such features must have arisen through lots of intermediates and, moreover, that these intermediate structures may once have served different functions from what we see today. The crystalline proteins that make up the lens of the eye, for example, are related to those that serve enzymatic functions unrelated to vision. So, the theory goes, evolution borrowed an existing protein and used it for a new function.

"Over time," Lenski said, "an old structure could be tweaked here and there to improve it for its new function, and that's a lot easier than inventing something entirely new."

That's where ALife sheds light.

"Darwinian evolution is a process that doesn't specify exactly how the evolving information is coded," says Adami, who leads the Digital Life Laboratory at Caltech. "It affects DNA and computer code in much the same way, which allows us to study evolution in this electronic medium."

Many computer scientists and engineers are now using processes based on principles of genetics and evolution to solve complex problems, design working robots, and more. Ofria says that "we can then apply these concepts when trying to decide how best to solve computational problems."

"Evolutionary design," says Pennock, "can often solve problems better than we can using our own intelligence."

[donh]

Yup. Essentially, atheists know that God exists but hate to admit it.

prove it.

[donh]

Each time I say abiogenesis is impossible you argue with me.

For me to argue with you, you'd have to actually pay attention. I have yet to see your proof, or even your hint, as to why I should suppose prokariotes, which is what you are describing, could only have sprung up out of organic junk instantaneously.

[donh]

Well, since you claim to have read the book, kindly tell us exactly what particles of matter can generate turing machines. This I need to see, I can use a good laugh.

Where did I claim that, exactly? Please show me before we go on--I'm pretty tired of you making stuff up that I've supposedly said and then putting on these supercelious tap-dancing shows when asked to produce the evidence from my own mouth.

In other words donh, I am asking you to back up your statement.

You have postulated that abiogenesis is categorically impossible. Yours the claim that needs backing up. My claim is quite a modest, but apposite one: that you haven't demonstrated compelling--even vaguely tempting, really--evidence, much less proof, of these ambitious universal claims.

[HalfFull]

btt

[donh]

.In what manner is this a relevant response to the extract quoted?

Seems quite obvious to me that matter cannot reason

Which is, to repeat myself, since I was apparently ignored, relevant to what I said how?

[donh]

No need to get angry. I'm only pointing out an inconsistency in your thinking. That doesn't translate to an attack or a threat or a put-down.

Let's get our emotional responses in order here. The problem is not anger, it is annoyance over what looks like obtuseness out of refusal to think about what is being read. As, for example, when you come right back with the same lazy, unresponsive misstatement of your deponents position, as you have just now done.

[donh]

But the idea of evolution came after faith and religion. So the disproving is up to your side.

Apparently, you do not understand what the word "faith" means. "Faith" is the proud claim that you haven't proved diddly. Like science, I've no interest in the saturday night grudge match you've arranged between God and science. Science doesn't care about God, one way or another--God is not within science's domain of competence.

[donh]

My argument against evolution is concicely stated in post#1265 just above.

So, apparently, you don't understand what "concise" means either. You have simply found an incredibly longwinded way to reject my suggestion that evolution did not always work the way it works in prokariotes. So it is just a restatement of your basic theme--that prokariotes had to leap into existence out of junk. This is the entire basis of your argument, absurd as it is, and you have simply window-dressed it to make it look more impressive than it is. It is likely that the rules governing meat machines are not going the be the rules for what went on before meat machines. My supposed contradiction is simply the fruit of me being willing to talk about pre- and post- meat machine paradigms--compounded by your inability to hone in on the precise details of any argument, busy as you are re-arranging your canned lecture yet again.

Pick and choose what you are arguing about and we can discuss that. Stop trying to purposely confuse the issues by saying that the argument against evolution does not apply to abiogenesis and that the argument against abiogenesis does not apply to evolution.

Take 5 seconds from your busy schedule to notice what I am actually arguing about regarding there being both a pre- and post- meat machine era, with distinct rules, and I'll consider discussing this further. Never mind agreeing or disagreeing with me--lets just see if you even understand.

[freeper4u]

freeper4u: "Programming a robot (or simulation of a robot) to walk in a real physical environment is hard problem"

AndrewC: "Don't you see the contradiction in your statement and how it relates to the rest of your statement?"

No.

[AndrewC]

No.

A simulation does not walk in a real physical environment. Programs are concepts as demonstrated in all of the virtual creatures that were "created" in the link you cited. IOW nothing that you presented as evidence was a material object.

[freeper4u]

Andrew,

In this example, the point I set out to make, and I think it was proved in the paper is that many problems that would be hard to solve by direct computation can be easily solved indirectly by genetic algorithms.

No simulation will ever be as good as the real world. But you can still learn a lot about the real world from a simulation.

Of course there are a zillion other hard problems that real creatures need to solve in the real world, but the Sims simulation effectively singled out a few of them and successfully showed they can be addressed quite effectively by genetic algorithms instead of a solution arrived at by direct design.

[colorado tanker]

No simulation will ever be as good as the real world. But you can still learn a lot about the real world from a simulation.

You mean like the global warming models??

[AndrewC]

but the Sims simulation effectively singled out a few of them and successfully showed they can be addressed quite effectively by genetic algorithms instead of a solution arrived at by direct design.

What was demonstrated was virtual. Real joints, with real motive elements requiring real energy sources and real raw materials occuring in a hostile environment was not demonstrated.

[freeper4u]

Andrew,

All that was shown in the Sims simulation was the question of whether or not it is possible for any kind of sensors, motors, and joints to be represented in a binary string, whether a sequence of binary operations on that string can improve the operations of those sensors, motors, and joints in a given environment for a given goal, and whether that binary string is sufficiently small enough that it does not rule out the possibility for a real DNA to represent the same information encoded by the simulation.

Even though that question is far different from the one of: "did real creatures evolve" and even "can real creatures evolve", it is still a very important question, and the result was positive, meaning that the next set of questions can be asked in the context of the hypothesis proved by this first experiment.

That is how science marches forward. It will never prove real creatures did evolve, but successive successful experiments show that it is more likely that evolution alone cannot be precluded as a possible explanation for the genesis of life on earth.

[AndrewC]

All that was shown in the Sims simulation was the question of whether or not it is possible for any kind of sensors, motors, and joints to be represented in a binary string, whether a sequence of binary operations on that string can improve the operations of those sensors, motors, and joints in a given environment for a given goal, and whether that binary string is sufficiently small enough that it does not rule out the possibility for a real DNA to represent the same information encoded by the simulation.

That is all well and good, but we represent things all of the time when we draw it, write about it, speak about it or think about it. We can even imagine improving it. The fact is everything that man has built started out as a concept. Even a rock used as a weapon had to be conceived even if it was initially picked up to be used as food.

As for DNA coding, anything can be used to represent something else because representation is a concept(written while moving salt shaker to the left and stating this is New York). Eyeblinks were used to send a message by a POW. You want to code the human genome by drumbeats, it can be done, but you better have a lot of time to waste.

You want to use DNA to code a program? That is easy. With 4 bases ,

1. Adenilne (A)
2. Cytosine (C)
3. Guanine (G)
4. Thymine (T) or Uracil (U)

you have a four level code. You can match the bases to 2 bits in this manner for an example.

• A = 00 (0)
• C = 01 (1)
• G = 10 (2)
• T = 11 (3)

Pentium instruction INT 7 has a hex value of CD07 or in binary

1100110100000111

In DNA code mapped as above this would be

TATCAACT

As you can see each byte takes up 4 bases, so that a DNA the size of the Human genome could code a program of about 750 megabytes. In this day and age it may seem like a trivial amount of data since it will easily fit on a DVD. However, going through all of the progrms that could be written in that amount of memory would take a very long time. Here is a previous calculation giving an indication of the information that can be coded by the DNA.

The number of different items represented by that 6 billion bit sequence as I stated before is 2^6000000000. In base 10 that would be 10^1806179974. Now there are 60 seconds/minute * 60 minutes/hour * 24 hours/day * 366 days/year * 15,000,000,000 years = 4.74336E+17 (or 10^17.67608609) seconds since the purported start of the universe. We will assume this full time is available to generate the sequences represented by the 6 billion bit sequence. The huge number represented by 10^1806179974 breaks all my calculators so I must use logs. If we divide the number of items (10^1806179974) by the time available (10^17.67608609) we end up with 10^1806179956 combinations/second. Clearly that is a huge amount of "information".

[AndrewC]

Proofread better

Adenilne = Adenoid Adenine

[Michael121]

Well let me set you straight. I made a comment. Someone did not like it. I like an argument so I kept it going.

The thing is this is non winnable to either side. Not that one side will not be proven correct. But that neither side will convince the other in my lifetime. Once my life is over, I guess I shall then see if faith was warrented.

If not then I have lost nothing. Science is cool. But what I see is that politics get involved way too much and in a lot of cases it appears, whether or not it does in actuality, to influence results. I like pure science that catters not to political pressures and PCism. If the science is real then it as also true.

Whether the science is geo-, bio-, or not there is a truth. The problem is we have to wait for one conclusion with political overtones. Before someone without politics gives the real answers.

There was a reporter after Columbine, who wanted to see how easy it was to get a gun at shows or on the WEB. He went into this with certain expectations. That is he thought it would be so easy to obtain a gun. He was proved wrong. The only person who would actually sell him a gun wanted him to drive 100+ miles to meet and make the sale in person, at a gun shop which could do a background check.

One example but there are hundreds like this based on science. Arts. Politics. Archeology. etc...

Politics is needed, and funny. It does however get in the way too often.

[longshadow]

multi-spectral moron alert placemarker

[freeper4u]

Andrew,

The beauty of genetic algorithms is that the encoding format of a DNA string is independent from the process of evolution because the agent making changes to the string has no idea what it is doing.

As in real evolution, the Sims experiment used a random process to mate the genes of its virtual creatures, so it doesn't matter how the encoded creature are represented. That is, it is not necessary for anyone to "design" a particular representation that enables genetic algorithms to evolve a string that decodes to a creature that improves its results in the fitness test.

[Ichneumon]

[Why don't you answer the question?]

You have not provided the evidence for the assertion that the evolved circuit performs better than the patented circuit even in its modified form. Not one measurement, not one graph, not one number.

Sigh -- just how many new pointless "objections" are you going to pull out of your hat in an endless parade of lame excuses to avoid having to actually deal with the questions that have been put to you, and the issues raised?

You're not fooling anyone with this game-playing. Except for perhaps yourself.

But just to give you one less cheap excuse to avoid the issue, here's more information on the evolved circuit: "Evolving Inventions", John R. Koza, Martin A. Keane and Matthew J. Streeter, "Scientific American", Feb. 2003. For a more technical treatment, see: Matthew J. Streeter, Martin A. Keane, John R. Koza: Routine Duplication of Post-2000 Patented Inventions by Means of Genetic Programming. EuroGP 2002: 26-36.

So what's your excuse going to be *now*, Mr. Evasive?

While you're at it, you might want to take a gander at the following and explain why evolution doesn't actually work in *these* research projects either, even though it certainly seems to perform just fine:

• Justin Balthrop, Fernando Esponda, Stephanie Forrest, Matthew Glickman: Coverage and Generalization in an Artificial Immune System. 3-10
• Mauro Birattari, Thomas Stützle, Luis Paquete, Klaus Varrentrapp: A Racing Algorithm for Configuring Metaheuristics. 11-18
• T. M. Blackwell, Peter J. Bentley: Dynamic Search With Charged Swarms. 19-26
• Christian Blum: Ant Colony Optimization For The Edge-weighted k-cardinality Tree Problem. 27-34
• Christian Blum, Michael Sampels, Mark Zlochin: On A Particularity In Model-based Search. 35-42
• Thang Nguyen Bui, Lisa C. Strite: An Ant System Algorithm For Graph Bisection. 43-51
• John A. Bullinaria: The Evolution Of Variable Learning Rates. 52-59
• Alex Conradie, Risto Miikkulainen, Christiaan Aldrich: Adaptive Control Utilising Neural Swarming. 60-67
• R. J. W. Hodgson: Partical Swarm Optimization Applied To The Atomic Cluster Optimization Problem. 68-73
• Christian Keber, Matthias G. Schuster: Option Valuation With Generalized Ant Programming. 74-81
• Nicole P. Leahy: Effects Of Agent Representation On The Behavior Of A Non-reciprocal Cooperation Game. 82-87
• Suihong Liang, A. Nur Zincir-Heywood, Malcolm I. Heywood: Intelligent Packets For Dynamic Network Routing Using Distributed Genetic Algorithm. 88-96
• Keith E. Mathias, Jason S. Byassee: Agent Support Of Genetic Search In An Immunological Model Of Sparse Distributed Memory. 97-104
• Daniel Merkle, Martin Middendorf: Studies On The Dynamics Of Ant Colony Optimization Algorithms. 105-112
• Kenneth O. Stanley, Risto Miikkulainen: Continual Coevolution Through Complexification. 113-120
• Keiki Takadama, Yutaka L. Suematsu, Norberto Eiji Nawa, Katsunori Shimohara: Cross-validation In Multiagent-based Simulation: Analyzing Evolutionary Bargaining Agents. 121-128
• Shin Ando, Hitoshi Iba: Ant Algorithm For Construction Of Evolutionary Tree. 131
• Josh C. Bongard, Rolf Pfeifer: Behavioural Selection Pressure Generates Hierarchical Genetic Regulatory Networks. 132
• Mariusz Boryczka, Zbigniew J. Czech: Solving Approximation Problems By Ant Colony Programming. 133
• Mathieu S. Capcarrère: Evolution Of Asynchronous Cellular Automata: Finding The Good Compromise. 134
• Sanjoy Das, Shekhar V. Gosavi, William H. Hsu, Shilpa A. Vaze: An Ant Colony Approach For The Steiner Tree Problem. 135
• Tom Lenaerts, A. Defaweux, Piet van Remortel, Bernard Manderick: An Individual-based Approach To Multi-level Selection. 136
• Olivier Pelletier, André Weimerskirch: Algorithmic Self-assembly Of DNA Tiles And Its Application To Cryptanalysis. 139-146
• Bernard Yurke, Friedrich C. Simmel: A DNA-based Three-state Device. 147-152
• Julie Beaulieu, Christian Gagné, Marc Parizeau: Lens System Design And Re-engineering With Evolutionary Algorithms. 155-162
• John C. Gallagher, Saranyan Vigraham: A Modified Compact Genetic Algorithm For The Intrinsic Evolution Of Continuous Time Recurrent Neural Networks. 163-170
• Morten Hartmann, Frode Eskelund, Pauline C. Haddow, Julian F. Miller: Evolving Fault Tolerance On An Unreliable Technology Platform. 171-177
• Uwe Tangen: An Evolvable Micro-controller Or What's New About Mutations? 178-186
• Nattee Niparnan, Prabhas Chongstitvatana: An Improved Genetic Algorithm For The Inference Of Finite State Machine. 189
• Brian K. Beachkofski, Gary B. Lamont: Evolutionary Programming Based Stratified Design Space Sampling. 193-200
• Carlos A. Coello Coello, Ricardo Landa Becerra: Adding Knowledge And Efficient Data Structures To Evolutionary Programming: A Cultural Algorithm For Constrained Optimization. 201-209
• Mikhail A. Semenov: Convergence Velocity Of Evolutionary Algorithm With Self-adaptation. 210-213
• Man Leung Wong, Shing Yan Lee, Kwong-Sak Leung: A Hybrid Data Mining Approach To Discover Bayesian Networks Using Evolutionary Programming. 214-222
• Yong Liu, Xin Yao: Search Step Size Control In Fast Evolutionary Programming. 225
• John DeLaurentis, Lauren Ferguson, William E. Hart: On The Convergence Properties Of A Simple Self-adaptive Evolutionary Algorithm. 229-237
• Thomas Jansen, Kenneth A. De Jong: An Analysis Of The Role Of Offspring Population Size In EAs. 238-246
• Tatsuya Okabe, Yaochu Jin, Bernhard Sendhoff: On The Dynamics Of Evolutionary Multi-objective Optimization. 247-256
• David J. John: Co-evolution With The Bierwirth-Mattfeld Hybrid Scheduler. 259
• Uwe Aickelin, Larry Bull: Partnering Strategies For Fitness Evaluation In A Pyramidal Evolutionary Algrorithm. 263-270
• Laura A. Albert, David E. Goldberg: Efficient Discretization Scheduling In Multiple Dimensions. 271-278
• Matthew Alden, Aard-Jan van Kesteren, Risto Miikkulainen: Eugenic Evolution Utilizing A Domain Model. 279-286
• Helio J. C. Barbosa, Afonso C. C. Lemonge: An Adaptive Penalty Scheme In Genetic Algorithms For Constrained Optimiazation Problems. 287-294
• Jason S. Byassee, Keith E. Mathias: Expediting Genetic Search With Dynamic Memory. 295-302
• Erick Cantú-Paz: Feature Subset Selection By Estimation Of Distribution Algorithms. 303-310
• Erick Cantú-Paz: On Random Numbers And The Performance Of Genetic Algorithms. 311-318
• Jian-Hung Chen, David E. Goldberg, Shinn-Ying Ho, Kumara Sastry: Fitness Inheritance In Multi-objective Optimization. 319-326
• Sung-Soon Choi, Byung Ro Moon: Isomorphism, Normalization, And A Genetic Algorithm For Sorting Network Optimization. 327-334
• Sung-Soon Choi, Byung Ro Moon: More Effective Genetic Search For The Sorting Network Problem. 335-342
• Federico Divina, Elena Marchiori: Evolutionary Concept Learning. 343-350
• César A. Galindo-Legaria, Florian Waas: The Effect Of Cost Distributions On Evolutionary Optimization Algorithms. 351-358
• Simon M. Garrett, Joanne H. Walker: Combining Evolutionary And Non-evolutionary Methods In Tracking Dynamic Global Optima. 359-366
• William A. Greene: A Genetic Algorithm With Self-distancing Bits But No Overt Linkage. 367-374
• Steffen G. Hohmann, Johannes Schemmel, Felix Schürmann, Karlheinz Meier: Exploring The Parameter Space Of A Genetic Algorithm For Training An Analog Neural Network. 375-382
• William H. Hsu, Haipeng Guo, Benjamin B. Perry, Julie A. Stilson: A Permutation Genetic Algorithm For Variable Ordering In Learning Bayesian Networks From Data. 383-390
• Michael Hüsken, Christian Igel: Balancing Learning And Evolution. 391-398
• Hisao Ishibuchi, Takashi Yamamoto: Fuzzy Rule Selection By Data Mining Criteria And Genetic Algorithms. 399-406
• Jung Hwan Kim, Byung Ro Moon: Neuron Reordering For Better Neuro-genetic Hybrids. 407-414
• Steven O. Kimbrough, Ming Lu, David Harlan Wood, Dong-Jun Wu: Exploring A Two-market Genetic Algorithm. 415-422
• Thomas E. Koch, Andreas Zell: MOCS: Multi-objective Clustering Selection Evolutionary Algorithm. 423-430
• Sujay V. Kumar, S. Ranji Ranjithan: Evaluation Of The Constraint Method-based Evolutionary Algorithm (CMEA) For A Three-objective Problem. 431-438
• Marco Laumanns, Lothar Thiele, Eckart Zitzler, Kalyanmoy Deb: Archiving With Guaranteed Convergence And Diversity In Multi-objective Optimization. 439-447
• Yong Liu, Xin Yao: Maintaining Population Diversity By Minimizing Mutual Information. 448-455
• Jiangming Mao, Kotaro Hirasawa, Jinglu Hu, Junichi Murata: Increasing Robustness Of Genetic Algorithm. 456-462
• Anil Menon: The Point Of Point Crossover: Shuffling To Randomness. 463-471
• Peter Merz: A Comparison Of Memetic Recombination Operators For The Traveling Salesman Problem. 472-479
• Mark M. Meyesenburg, Daniel Hoelting, Duane McElvain, James A. Foster: How Random Generator Quality Impacts GA Performance. 480-487
• Miguel Nicolau, Conor Ryan: LINKGAUGE: Tackling Hard Deceptive Problems With A New Linkage Learning Genetic Algorithm. 488-494
• Gabriela Ochoa: Setting The Mutation Rate: Scope And Limitations Of The 1/L Heuristic. 495-502
• Liviu Panait, Sean Luke: A Comparison Of Two Competitive Fitness Functions. 503-511
• Martin Pelikan, David E. Goldberg, Shigeyoshi Tsutsui: Combining The Strengths Of Bayesian Optimization Algorithm And Adaptive Evolution Strategies. 512-519
• Robin C. Purshouse, Peter J. Fleming: Why Use Elitism And Sharing In A Multi-objective Genetic Algorithm? 520-527
• Kumara Sastry, David E. Goldberg: Genetic Algorithms, Efficiency Enhancement, And Deciding Well With Differing Fitness Variances. 528-535
• Kumara Sastry, David E. Goldberg: Genetic Algorithms, Efficiency Enhancement, And Deciding Well With Differing Fitness Bias Values. 536-543
• Dong-il Seo, Byung Ro Moon: Voronoi Quantizied Crossover For Traveling Salesman Problem. 544-552
• Hiroshi Someya, Masayuki Yamamura: Robust Evolutionary Algorithms With Toroidal Search Space Conversion For Function Optimization. 553-560
• Alexander V. Spirov, Alexander B. Kazansky: Jumping Genes-mutators Can Rise Efficacy Of Evolutionary Search. 561-568
• Kenneth O. Stanley, Risto Miikkulainen: Efficient Reinforcement Learning Through Evolving Neural Network Topologies. 569-577
• Christopher R. Stephens, Riccardo Poli, Alden H. Wright, Jonathan E. Rowe: Exact Results From A Coarse Grained Formulation Of The Dynamics Of Variable-length Genetic Algorithms. 578-585
• Christopher Stone, Jim Smith: Strategy Parameter Variety In Self-adaptation Of Mutation Rates. 586-593
• Hal Stringer, Annie S. Wu: A Simple Method For Detecting Domino Convergence And Identifying Salient Genes Within A Genetic Algorithm. 594-601
• Ashutosh Tiwari, Rajkumar Roy: Variable Dependence Interaction And Multi-objective Optimisation. 602-609
• Huai-Kuang Tsai, Jinn-Moon Yang, Cheng-Yan Kao: Applying Genetic Algorithms To Finding The Optimal Gene Order In Displaying The Microarray Data. 610-617
• Alexander F. Tulai, Franz Oppacher: Combining Competitive And Cooperative Coevolution For Training Cascade Neural Networks. 618-625
• Clarissa Van Hoyweghen, David E. Goldberg, Bart Naudts: From Twomax To The Ising Model: Easy And Hard Symmetrical Problems. 626-633
• Dana Vrajitoru: Simulating Gender Separation With Genetic Algorithms. 634-641
• Alden H. Wright, Jonathan E. Rowe, Riccardo Poli, Christopher R. Stephens: A Fixed Point Analysis Of A Gene Pool GA With Mutation. 642-649
• Shengxiang Yang: Adaptive Non-uniform Crossover Based On Statistics For Genetic Algorithms. 650-657
• Zhong-Yao Zhu, Kwong-Sak Leung: An Enhanced Annealing Genetic Algorithm For Multi-objective Optimization Problems. 658-665
• Eckart Zitzler, Marco Laumanns, Lothar Thiele, Carlos M. Fonseca, Viviane Grunert da Fonseca: Why Quality Assessment Of Multiobjective Optimizers Is Difficult. 666-674
• Jaume Bacardit, Josep Maria Garrell i Guiu: Evolution Of Adaptive Discretization Intervals For A Rule-based Genetic Learning System. 677
• Dirk Devogelaere, Marcel Rijckaert: Influences Of Clustering Modifications On The Performatnce Of The Genetic Algorithm Driven Clustering Algorithm. 678
• A. Sima Etaner-Uyar, A. Emre Harmanci: Preserving Diversity In Changing Environments Through Diploidy With Adaptive Dominance. 679
• William H. Hsu, Cecil P. Schmidt, James A. Louis: Genetic Algorithm Wrappers For Feature Subset Selection In Supervised Inductive Learning. 680
• Chien-Geng Huang: A Study Of Fitness Proportional Mate Selection Schemes In Genetic Algorithms. 681
• Chien-Feng Huang: A Markov Chain Analysis Of Fitness Proportional Mate Selection Schemes In Genetic Algorithm. 682
• Yaochu Jin, Bernhard Sendhoff: Incorporation Of Fuzzy Preferences Into Evolutionary Multiobjective Optimization. 683
• Nazan Khan, David E. Goldberg, Martin Pelikan: Multiple-objective Bayesian Optimization Algorithm. 684
• Jong-Pil Kim, Byung Ro Moon: A Hybrid Genetic Search For Circuit Bipartitioning. 685
• Yong-Hyuk Kim, Byung Ro Moon: Visualization Of The Fitness Landscape, A Steady-state Genetic Search, And Schema Traces. 686
• Kosmas Knödler, Jan Poland, Andreas Zell, Alexander Mitterer: Memetic Algorithms For Combinatorial Optimization Problems In The Calibration Of Modern Combustion Engines. 687
• Alexander Kosorukoff: Using Incremental Evaluation And Adaptive Choice Of Operators In A Genetic Algorithm. 688
• Seung-Kyu Lee, Dong-il Seo, Byung Ro Moon: A Hybrid Genetic Algorithm For Optimal Hexagonal Tortoise Problem. 689
• Penousal Machado, Jorge Tavares, Francisco B. Pereira, Ernesto Costa: Vehicle Routing Problem: Doing It The Evolutionary Way. 690
• Mark M. Meyesenburg, Dan Hoelting, Duane McElvain, James A. Foster: A Genetic Algorithm-specific Test Of Random Generator Quality. 691
• James E. Pettinger, Richard M. Everson: Controlling Genetic Algorithms With Reinforcement Learning. 692
• Oclair Prado, Fernando J. Von Zuben: An Integrated System For Phylogenetic Inference Using Evolutionary Algorithms. 693
• Robert S. Roos, Tiffany Bennett, Jennifer Hannon, Elizabeth Zehner: A Genetic Algorithm For Improved Shellsort Sequences. 694
• Franz Rothlauf: The Influence Of Binary Representations Of Integers On The Performance Of Selectorecombinative Genetic Algorithms. 695
• Ivan Sekaj, Martin Foltin, Michal Gonos: Genetic Algorithm Based Adaptive Control Of An Electromechanical MIMO System. 696
• Anabela Simões, Ernesto Costa: Parametric Study To Enhance The Genetic Algorithm's Performance When Using Transformation. 697
• Anabela Simões, Ernesto Costa: Using GAs To Deal With Dynamic Environments: A Comparative Study Of Several Approaches Based On Promoting Diversity. 698
• Abhishek Singh, David E. Goldberg, Ying-Ping Chen: Modified Linkage Learning Genetic Algorithm For Difficult Non-stationary Problems. 699
• Tapio Tyni, Jari Ylinen: Bi-directional Circular Linked Lists In Fitness Caching. 700
• Takanori Ueda, Nobuto Koga, Isao Ono, Masahiro Okamoto: Application Of Numerical Optimization Technique Based On Real-coded Genetic Algorithm To Inverse Problem In Biochemical Systems. 701
• Shinya Watanabe, Tomoyuki Hiroyasu, Mitsunori Miki: LCGA: Local Cultivation Genetic Algorithm For Multi-objective Optimization Problems. 702
• Annie S. Wu, Ivan Garibay: The Proportional Genetic Algorithm Representation. 703
• Tina Yu, Julian F. Miller: Climbing Unimodal Landscapes With Neutrality: A Case Study Of The One-max Problem. 704
• Atif Azad, Conor Ryan, Mark E. Burke, Ali R. Ansari: A Re-examination Of The Cart Centering Problem Using The Chorus System. 707-715
• Raymond Burke, Steven Gustafson, Graham Kendall: A Survey And Analysis Of Diversity Measures In Genetic Programming. 716-723
• Manuel Clergue, Philippe Collard, Marco Tomassini, Leonardo Vanneschi: Fitness Distance Correlation And Problem Difficulty For Genetic Programming. 724-732
• Raphael Crawford-Marks, Lee Spector: Size Control Via Size Fair Genetic Operators In The PushGP Genetic Programming System. 733-739
• R. Groß, K. Albrecht, Wolfgang Kantschik, Wolfgang Banzhaf: Evolving Chess Playing Programs. 740-747
• Lutz H. Hamel: Breeding Algebraic Structures - An Evolutionary Approach To Inductive Equational Logic Programming. 748-755
• Daniel Howard, Simon C. Roberts, Conor Ryan: Machine Vision: Exploring Context With Genetic Programming. 756-763
• William H. Hsu, Steven M. Gustafson: Genetic Programming And Multi-agent Layered Learning By Reinforcements. 764-771
• Jianjun Hu, Erik D. Goodman, Kisung Seo, Min Pei: Adaptive Hierarchical Fair Competition (AHFC) Model For Parallel Evolutionary Algorithms. 772-779
• Jianjun Hu, Kisung Seo, Shaobo Li, Zhun Fan, Ronald C. Rosenberg, Erik D. Goodman: Structure Fitness Sharing (SFS) For Evolutionary Design By Genetic Programming. 780-787
• Hitoshi Iba, Erina Sakamoto: Inference Of Differential Equation Moels By Genetic Programming. 788-795
• Kosuke Imamura, Robert B. Heckendorn, Terence Soule, James A. Foster: Abstention Reduces Errors - decision Abstaining N-version Genetic Programming. 796-803
• Jeremy Kubica, Eleanor G. Rieffel: Collaborating With A Genetic Programming System To Generate Modular Robotic Code. 804-811
• William B. Langdon: Convergence Rates For The Distribution Of Program Outputs. 812-819
• Sean Luke, Liviu Panait: Is The Perfect The Enemy Of The Good? 820-828
• Sean Luke, Liviu Panait: Lexicographic Parsimony Pressure. 829-836
• Peter Martin: An Analysis Of Random Number Generators For A Hardware Implementation Of Genetic Programming Using FPGAs And Handel-C. 837-844
• Peter Martin, Riccardo Poli: Crossover Operators For A Hardware Implementation Of GP Using FPGAs And Handel-C. 845-852
• Nicholas Freitag McPhee, Riccardo Poli: Using Schema Theory To Explore Interactions Of Multiple Operators. 853-860
• Johan Parent, Ann Nowé: Evolving Compression Preprocessors With Genetic Programming. 861-867
• Riccardo Poli, Christopher R. Stephens, Alden H. Wright, Jonathan E. Rowe: On The Search Biases Of Homologuous Crossover In Linear Genetic Programming And Variable-length Genetic Algorithms. 868-876
• Matthew J. Streeter, Martin A. Keane, John R. Koza: Iterative Refinement Of Computational Circuits Using Genetic Programming. 877-884
• Jason Cooper, Chris Hinde: Comparison Of Evolving Against Peers And Fixed Opponents Using Corewars. 887
• Christian Gagné, Marc Parizeau: Open BEAGLE: A New C++ Evolutionary Computation Framework. 888
• Mario Giacobini, Marco Tomassini, Leonardo Vanneschi: How Statistics Can Help In Limiting The Number Of Fitness Cases In Genetic Programming. 889
• Hironobu Katagiri, Kotaro Hirasawa, Jinglu Hu, Junichi Murata: A New Model To Realize Variable Size Genetic Network Programming. 890
• Emin Erkan Korkmaz, Göktürk Ücoluk: Controlling The Genetic Programming Search. 891
• Carlos Oliver-Morales, Katya Rodríguez Vázquez: MB GP In Modelling And Prediction. 892
• Ronald Olson, Brock Wilcox: Self-improvement For The ADATE Automatic Programming System. 893
• Mark S. Withall, Chris J. Hinde, Roger G. Stone: Evolving Readable Perl. 894
• Larry Bull: Lookahead And Latent Learning In ZCS. 897-904
• Larry Bull, Toby O'Hara: Accuracy-based Neuro And Neuro-fuzzy Classifier Systems. 905-911
• Martin Danek, Robert E. Smith: XCS Applied To Mapping FPGA Architectures. 912-919
• Chunsheng Fu, Lawrence Davis: A Modified Classifier System Compaction Algorithm. 920-925
• Samuel Landau, Sébastien Picault, Oliver Sigaud, Pierre Gérard: A Comparison Between ATNoSFERES And XCSM. 926-933
• Xavier Llorà, Josep Maria Garrell i Guiu: Coevolving Different Knowledge Representations With Fine-grained Parallel Learning Classifier Systems. 934-941
• Peter Ross, Sonia Schulenburg, Javier G. Marín-Blázquez, Emma Hart: Hyper-heuristics: Learning To Combine Simple Heuristics In Bin-packing Problems. 942-948
• Larry Bull, Dave Wyatt, Ian C. Parmee: Towards The Use Of XCS In Interactive Evolutionary Design. 951
• Gabriella Kókai, Zoltán Tóth, Szilvia Zvada: An Experimental Comparison Of Genetic And Classical Concept Learning Methods. 952
• Filippo Neri: Cooperative Concept Learning By Means Of A Distributed GA. 953-956
• Marc Ebner, Hans-Georg Breunig, Jürgen Albert: On The Use Of Negative Selection In An Artificial Immune System. 957-964
• Alexander Kosorukoff, David E. Goldberg: Evolutionary Computation As A Form Of Organization. 965-972
• Hassan Z. Masum, Steffen Christensen, Franz Oppacher: The Turing Ratio: Metrics For Open-ended Tasks. 973-980
• Xiaotong Wang, Lawrence Davis, Chunsheng Fu: Genetic Algorithms And Fine-grained Topologies For Optimization. 981-988
• Maribel García Arenas, Brad Dolin, Juan J. Merelo Guervós, Pedro A. Castillo Valdivieso, Ignacio Fernández De Viana, Marc Schoenauer: JEO: Java Evolving Objects. 991-994
• L. Barone, L. While, P. Hingston: Designing Crushers With A Multi-objective Evolutionary Algorithm. 995-1002
• Bir Bhanu, Yingqiang Lin: Learning Composite Operators For Object Detection. 1003-1010
• Anthony Brabazon, Michael O'Neill, Robin Matthews, Conor Ryan: Grammatical Evolution And Corporate Failure Prediction. 1011-1018
• Erick Cantú-Paz, Chandrika Kamath: Evolving Neural Networks For The Classification Of Galaxies. 1019-1026
• Robert Carr, William Hart, Natalio Krasnogor, Jonathan Hirst, Edmund K. Burke, James Smith: Alignment Of Protein Structures With A Memetic Evolutionary Algorithm. 1027-1034
• Deborah R. Carvalho, Alex Alves Freitas: A Genetic Algorithm With Sequential Niching For Discovering Small-disjunct Rules. 1035-1042
• Flor A. Castillo, Ken A. Marshall, James L. Green, Arthur K. Kordon: Symbolic Regression In Design Of Experiments: A Case Study With Linearizing Transformations. 1043-1047
• Ping Chen, Zhaohui Fu, Ping Chen, Andrew Lim: Using Genetic Algorithms To Solve The Yard Allocation Problem. 1049-1056
• Clare Bates Congdon: Gaphyl: An Evolutionary Algorithms Approach For The Study Of Natural Evolution. 1057-1064
• Anthony Di Pietro, Lyndon While, Luigi Barone: Learning In RoboCup Keepaway Using Evolutionary Algorithms. 1065-1072
• Zhun Fan, Kisung Seo, Ronald C. Rosenberg, Jianjun Hu, Erik D. Goodman: Exploring Multiple Design Topologies Using Genetic Programming And Bond Graphs. 1073-1080
• Fabio A. Gonzalez, Dipankar Dasgupta: An Imunogenetic Technique To Detect Anomalies In Network Traffic. 1081-1088
• Karim Hamza, Haitham Mahmoud, Kazuhiro Saitou: Design Optimization Of N-shaped Roof Trusses. 1089-1096
• Daniel Howard, Simon C. Roberts: Application Of Genetic Programming To Motorway Traffic Modelling. 1097-1104
• Yaochu Jin, Bernhard Sendhoff: Fitness Approximation In Evolutionary Computation - a Survey. 1105-1112
• Li-Shan Kang, Zhou Kang, Yan Li, Hugo de Garis: A Two Level Evolutionary Modeling System For Financial Data. 1113-1118
• Yung-Keun Kwon, Sung-Deok Hong, Byung Ro Moon: A Genetic Hybrid For Critical Heat Flux Function Approximation. 1119-1125
• Sang-Yon Lee, Sung-Soon Choi, Byung Ro Moon: Search Improvement By Genetic Algorithms With A Semiotic Network. 1126-1132
• Derek S. Linden: Antenna Design Using Genetic Algorithm. 1133-1140
• David Montana, Talib Hussain, Tushar Saxena: Adaptive Reconfiguration Of Data Networks Using Genetic Algorithms. 1141-1149
• Jason H. Moore, Lance W. Hahn, Marylyn D. Ritchie, Tricia A. Thornton, Bill C. White: Application Of Genetic Algorithms To The Discovery Of Complex Models For Simulation Studies In Human Genetics. 1150-1155
• Boris Naujoks, Werner Haase, Jörg Ziegenhirt, Thomas Bäck: Multi Objective Airfoil Design Using Single Parent Populations. 1156-1163
• V. Oduguwa, R. Roy: Multi-objective Optimisation Of Rolling Rod Product Design Using Meta-modelling Approach. 1164-1171
• Eun-Jong Park, Yong-Hyuk Kim, Byung Ro Moon: Genetic Search For Fixed Channel Assignment Problem With Limited Bandwidth. 1172-1179
• Khaled Rasheed, Swaroop Vattam, Xiao Ni: Comparison Of Methods For Using Reduced Models To Speed Up Design Optimization. 1180-1187
• Wesley Romão, Alex Alves Freitas, Roberto C. S. Pacheco: A Genetic Algorithm For Discovering Interesting Fuzzy Prediction Rules: Applications To Science And Technology Data. 1188-1195
• Brian J. Ross, Anthony G. Gualtieri, Frank Fueten, Paul Budkewitsch: Hyperspectral Image Analysis Using Genetic Programming. 1196-1203
• Yuji Sato: Voice Conversion Using Interactive Evolution Of Prosodic Control. 1204-1211
• J. David Schaffer, Lalitha Agnihotri, Nevenka Dimitrova, Thomas McGee, Sylvie Jeannin: Improving Digital Video Commercial Detectors With Genetic Algorithms. 1212-1218
• Ivan Tanev, Takashi Uozumi, Yoshiharu Morotome: An Application Service Provider Approach For Hybrid Evolutionary Algorithm-based Real-world Flexible Job Shop Scheduling Problem. 1219-1226
• Mark S. Voss, Xin Feng: A New Methodology For Emergent System Identification Using Particle Swarm Optimization (PSO) And The Group Mehtod Data Handling (GMDH). 1227-1232
• Joachim Wegener, Kerstin Buhr, Hartmut Pohlheim: Automatic Test Data Generation For Structural Testing Of Embedded Software Systems By Evolutionary Testing. 1233-1240
• Abdulnasser Younes, Hamada Ghenniwa, Shawki Areibi: An Adaptive Genetic Algorithm For Multi Objective Flexible Manufacturing Systems. 1241-1248
• Jian Zhang, Xiaohui Yuan, Bill P. Buckles: A Fast Evolution Strategies Based Approach To Image Registration. 1249-1256
• Sung-Soon Choi, Byung Ro Moon: Optimized Interest Metric Of Rules And One-to-one Marketing Using Connection Networks. 1259
• H. W. Chong, Sam Kwong: A Genetic Algorithm For Joint Optimization Of Spare Capacity And Delay In Self-healing Network. 1260
• L. Elliott, David B. Ingham, A. G. Kyne, N. S. Mera, M. Pourkashanian, C. W. Wilson: A Real Coded Genetic Algorithm For The Optimisation Of Reaction Rate Parameters For Chemical Kinetic Modelling In A Perfectly Stirred Reactor. 1261
• Sean L. Forman: Congressional Redistricting Using A TSP-based Genetic Algorithm. 1262
• Héctor Fernando Gómez García, Arturo González Vega, Arturo Hernández Aguirre, Carlos A. Coello Coello: Efficient Affine 2D-image Registration Using Evolutionary Strategies. 1263
• Marco César Goldbarg, Elizabeth Ferreira Gouvea, Francisco Dantas de Medeiros Neto: Piston Pump Mobile Unity Tour Problem: An Evolutionary View. 1264
• Alberto Gomez, David de la Fuente, Jose Parreño, Javier Puente: Using Genetic Algorithms To Optimize Guillotine Cutting Operations. 1265
• Alex C. H. Ho, Sam Kwong: Optimization Of CDMA Based Wireless System. 1266
• Zhou Ji, Dipankar Dasgupta: Modeling Convection Coefficients With Genetic Algorithms. 1267
• Otavio Larsen, Alex Alves Freitas, Júlio C. Nievola: Constructing X-of-n Attributes With A Genetic Algorithm. 1268
• Shouichi Matsui, Isamu Watanabe, Ken-ichi Tokoro: An Efficient Genetic Algorithm For Fixed Channel Assignment Problem With Limited Bandwidth Constraint. 1269
• Fernando E. B. Otero, Monique M. S. Silvia, Alex Alves Freitas: Genetic Programming For Attribute Construction In Data Mining. 1270
• Supiya Ujjin, Peter J. Bentley: Evolving Good Recommendations. 1271
• Edgar E. Vallejo, Fernando Ramos: Evolving Finite Automata With Two-dimensional Output For DNA Recognition And Visualization. 1272
• Andre Vogel, Marco Fischer, Tobias Teich: Real-world Shop Floor Scheduling By Ant Colony Optimization. 1273
• L. Wu, Walter D. Potter, K. Rasheed, H. Thistle, J. Ghent, D. Twardus, M. Teske: A Comparison of Genetic Algorithm Methods In Aerial Spray Deposition Management. 1274
• Xiaoming Yu, Alessandro Fin, Franco Fummi, Elizabeth M. Rudnick: Functional Test Generation For Digital Integrated Circuits Using A Genetic Algorithm. 1275
• Jesse B. Zydallis, Todd A. Sriver, Gary B. Lamont: Multiobjective Evolutionary Algorithm Approach For Solving Integer Based Optimization Problems. 1276
• Peter Augustsson, Krister Wolff, Peter Nordin: Creation Of A Learning, Flying Robot By Means Of Evolution. 1279-1285
• Gary B. Parker: Learning Area Coverage Using The Co-evolution Of Model Parameters. 1286-1294
• Karl Hedman, David Persson, Per Skoglund, Dan Wiklund, Krister Wolff, Peter Nordin: Sensing And Direction In Locomotion Learning With A Random Morphology Robot. 1297
• Jumpol Polvichai, Pradeep K. Khosla: Applying Dynamic Networks To Improve Learning Performances Of An Evolutionary Behavior Programming System For Mobile Robots In Dynamic Environments. 1298
• Hisao Ishibuchi, Tadashi Yoshida, Tadahiko Murata: Balance Between Genetic Search And Local Search In Hybrid Evolutionary Multi-criterion Optimization Algorithms. 1301-1308
• Soonchul Jung, Byung Ro Moon: A Hybrid Genetic Algorithm For The Vehicle Routing Problem With Time Windows. 1309-1316
• Marc Reimann, Michael Stummer, Karl Doerner: A Savings Based Ant System For The Vehicle Routing Problem. 1317-1326
• André Baresel, Harmen Sthamer, Michael Schmidt: Fitness Function Design To Improve Evolutionary Structural Testing. 1329-1336
• Leonardo Bottaci: Instrumenting Programs With Flag Variables For Test Data Search By Genetic Algorithms. 1337-1342
• Maria Cláudia Figueiredo Pereira Emer, Silva Regina Vergilio: GPTesT: A Testing Tool Based On Genetic Programming. 1343-1350
• Mark Harman, Robert M. Hierons, Mark Proctor: A New Representation And Crossover Operator For Search-based Optimization Of Software Modularization. 1351-1358
• Mark Harman, Lin Hu, Robert M. Hierons, André Baresel, Harmen Sthamer: Improving Evolutionary Testing By Flag Removal. 1359-1366
• Colin Kirsopp, Martin J. Shepperd, John Hart: Search Heuristics, Case-based Reasoning And Software Project Effort Prediction. 1367-1374
• Brian S. Mitchell, Spiros Mancoridis: Using Heuristic Search Techniques To Extract Design Abstractions From Source Code. 1375-1382
• Terry Van Belle, David H. Ackley: Code Factoring And The Evolution Of Evolvability. 1383-1390
• Hans-Gerhard Groß, Nikolas Mayer: Evolutionary Testing In Component-based Real-time System Construction. 1393

The above can be found in the publication: William B. Langdon, Erick Cantú-Paz, Keith E. Mathias, Rajkumar Roy, David Davis, Riccardo Poli, Karthik Balakrishnan, Vasant Honavar, Günter Rudolph, Joachim Wegener, Larry Bull, Mitchell A. Potter, Alan C. Schultz, J. F. Miller, E. Burke, Natasa Jonoska (Eds.): GECCO 2002: Proceedings of the Genetic and Evolutionary Computation Conference, New York, USA, 9-13 July 2002. Morgan Kaufmann 2002, ISBN 1-55860-878-8

Give it up, Andrew, you're just looking silly. There's a whole world of significant evolutionary research results out there that you can't just make go away by wishing hard enough, or being evasive enough.