J. Koza. Genetic evolution and co-evolution of computer programs. In J. D. F. Christopher G. Langton, Charles Taylor and S. Rasmussen, editors, Artificial Life II. Addison Wesley Publishing Company, Reading Mass., 1992.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....cellular automata. The idea was the possibility of observing the emergence of Lisp functions. There were two strong appeals for using this language. An empirical reason was that Lisp code had been used with great success in Koza s [1990] technique of genetic programming mentioned earlier (also in [Koza 1992], which I had read a preprint of) This technique is essentially a search method in the space of Lisp functions that finds a particular function to solve a predefined problem. Another suggestion one would get from the literature would be the use of an assembly like language, as in [Harvey 1991] ....

John R. Koza. Genetic evolution and co-evolution of computer programs. In C. G. Langton, J. D. Farmer, S. Rasmussen, and C. Taylor, editors, Artificial Life: Proceedings of the second workshop on Artificial Life, pages 603--629. Addison-Wesley, 1992.

....Program Evolution (PIPE) 19] PIPE is a novel technique for automatic program synthesis. It combines probability vector coding of program instructions [25, 26, 27] Population Based Incremental Learning [2] and tree coded programs like those used in some variants of Genetic Programming (GP) [7, 8, 12]. PIPE iteratively generates successive populations of functional programs according to an adaptive probability distribution over all possible programs. Each iteration it lets all programs play one soccer game; then the best program is used to refine the distribution. Thus PIPE stochastically ....

J. R. Koza. Genetic evolution and co-evolution of computer programs. In C. G. Langton, C. Taylor, J. D. Farmer, and S. Rasmussen, editors, Artificial Life II, pages 313--324. Addison Wesley Publishing Company, 1992.

....variety of different fields. Specifically, genetic programming has been successfully applied to problems such as . planning (e.g. navigating an artificial ant along a trail, developing a robotic action sequence that can stack an arbitrary initial configuration of blocks into a specified order) [Koza 1991], discovering control strategies for backing up a tractor trailer truck, centering a cart, and balancing a broom on a moving cart, discovering inverse kinematic equations to control the movement of a robot arm to a designated target point, evolution of a subsumption architecture for robotic ....

.... [Koza 1992c, Koza and Rice 1992b] emergent behavior (e.g. discovering a computer program which, when executed by all the ants in an ant colony, enables the ants to locate food, pick it up, carry it to the nest, and drop pheromones along the way so as to produce cooperative emergent behavior) [Koza 1991], classification and pattern recognition (e.g. distinguishing two intertwined spirals) generation of maximal entropy random numbers, induction of decision trees for classification, optimization problems (e.g. finding an optimal food foraging strategy for a lizard) Koza, Rice, and ....

Koza, John R. Genetic evolution and co-evolution of computer programs. In Langton, Christopher, Taylor, Charles, Farmer, J. Doyne, and Rasmussen, Steen (editors). Artificial Life II, SFI Studies in the Sciences of Complexity. Volume X. Redwood City, CA: Addison-Wesley 1991. Pages 603-629. 1991.

....an arti cial environment with two kinds of inhabitants: animals and plants. The animals are genetic which means that their physical and mental properties are determined by their chromosomes, and genetic operators, such as crossover and mutation, are used on the chromosomes to produce o spring. In (Koza, 1991), Rasmussen et al. 1991) Je erson et al. 1991) Koza, 1994) Ray, 1991) and (Skipper, 1991) the genetic information is based on programmable matter like executable code. The genetic information of the animals in our environment is based on dna and rna like structures (Schuster, 1991) The ....

Koza, J. R. (1991). Genetic evolution and co-evolution of computer programs. In Langdon, C. G., Taylor, C., Farmer, J. D., and Rasmussen, S., editors, Articial Life II, volume X. Addison-Wesley.

..... optimal control (e.g. centering a cart and balancing a broom on a moving cart in minimal time by applying a bang bang force to the cart [Koza and Keane 1990] and backing a tractor trailer truck to a loading dock [Koza 1992a] planning (e.g. navigating an artificial ant along a trail) [Koza 1991a] finding minimax strategies for games (e.g. differential pursuer evader games, discrete games in extensive form) by both evolution and co evolution [Koza 1991b] evolving robotic action plans in the style of the subsumption architecture (e.g. a wall following strategy for a robot with ....

.... and backing a tractor trailer truck to a loading dock [Koza 1992a] planning (e.g. navigating an artificial ant along a trail) Koza 1991a] finding minimax strategies for games (e.g. differential pursuer evader games, discrete games in extensive form) by both evolution and co evolution [Koza 1991b] evolving robotic action plans in the style of the subsumption architecture (e.g. a wall following strategy for a robot with sonar sensors in an irregular room) Koza 1992d] empirical discovery (e.g. rediscovering Kepler s Third Law, rediscovering the well known nonlinear econometric ....

[Article contains additional citation context not shown here]

Koza, John R. Genetic evolution and co-evolution of computer programs. In Langton, Christopher, Taylor, Charles, Farmer, J. Doyne, and Rasmussen, Steen (editors). Artificial Life II, SFI Studies in the Sciences of Complexity. Volume X. Redwood City, CA: Addison-Wesley 1991. Pages 603-629. 1991b.

....state and or history. This of course is quite unlike our own framework, where the dynamics of the atomic changes are the basis for change, and any statistical properties are derived from these. Work in population genetics (e.g. 2] 4] 20] and work in artificial life inspired by it (e.g. 21] [17]) is closer to ours in this sense. Here we have a set of individuals, each belonging to one of several types. The system evolves in generations ; in each generation the individual evolves in a way that is defined by its type and the environment (which includes the other agents) and at the end of ....

J. Koza. Genetic Evolution and Co-Evolution of Computer Programs. In C.G. Langton, C. Taylor, J.D. Farmer, and S. Rasmussen, editors, Artificial Life II. Addison-Wesley, 1992.

.... Red Queen [ Carroll, 1871 ] approach where the population must continually improve itself) A dynamic fitness function could be pre defined but dynamic GP fitness functions are often produced by co evolution [ Hillis, 1992; Angeline and Pollack, 1993; Angeline, 1993; Angeline and Pollack, 1994; Koza, 1991; Jannink, 1994; Reynolds, 1994; Ryan, 1995 ] Where it is felt certain characters will be required in the problem s solution the initial population and crossover can be controlled in order to ensure individuals within the population have these properties ( Langdon, 1995 ] and [ Langdon, 1996b ....

John R. Koza. Genetic evolution and co-evolution of computer programs. In Christopher Taylor Charles Langton, J. Doyne Farmer, and Steen Rasmussen, editors, Artificial Life II, volume X of SFI Studies in the Sciences of Complexity, pages 603--629. Addison-Wesley, Redwood City, CA, USA, 1991.

....smaller entities. This agglomerative operation is a highly simplified form of recombination (crossover) Another basic operation permits the emission of new entities into the sea. The approach used here bears some relation to both the genetic algorithm (Holland 1975, 1992) and genetic programming (Koza 1989, 1991, 1992) as illustrated on videotape (Koza and Rice 1991a, 1992) In genetic methods, the calculation of the fitness measure is usually an explicit calculation and the individuals in the population are passive. The fitness of each individual is measured by means of some explicit calculation and the ....

....highly simplified form of recombination (crossover) Another basic operation permits the emission of new entities into the sea. The approach used here bears some relation to both the genetic algorithm (Holland 1975, 1992) and genetic programming (Koza 1989, 1991, 1992) as illustrated on videotape (Koza and Rice 1991a, 1992) In genetic methods, the calculation of the fitness measure is usually an explicit calculation and the individuals in the population are passive. The fitness of each individual is measured by means of some explicit calculation and the controller of the genetic process applies various ....

Koza, John R. Genetic evolution and co-evolution of computer programs. In Langton, Christopher, Taylor, Charles, Farmer, J. Doyne, and Rasmussen, Steen (editors). Artificial Life II, SFI Studies in the Sciences of Complexity. Volume X. Redwood City, CA: Addison-Wesley. Pages 603--629. 1991.

....we have shown that computer programs can be genetically bred to solve a surprising variety of problems in many different areas [Koza 1989, 1990, 1992a] including . planning (e.g. navigating an artificial ant along a trail and developing a robotic plan for stacking blocks in to a desired order) [Koza 1989, 1991a] emergent behavior (e.g. discovering a computer program which, when executed by all the ants in an ant colony, enables the ants to locate food, pick it up, carry it to the nest, and drop pheromones along the way so as to produce cooperative emergent behavior) Koza 1991] evolution of ....

.... a desired order) Koza 1989, 1991a] emergent behavior (e.g. discovering a computer program which, when executed by all the ants in an ant colony, enables the ants to locate food, pick it up, carry it to the nest, and drop pheromones along the way so as to produce cooperative emergent behavior) [Koza 1991], evolution of subsumption (e.g. evolving a program for a wall following robot) Koza 1992b] machine learning of functions (e.g. learning the Boolean 11 multiplexer function) Koza 1991d] automatic programming (e.g. solving pairs of linear equations, solving quadratic equations for ....

[Article contains additional citation context not shown here]

Koza, John R. Genetic evolution and co-evolution of computer programs. In Langton, Christopher, Taylor, Charles, Farmer, J. Doyne, and Rasmussen, Steen (editors). Artificial Life II, SFI Studies in the Sciences of Complexity. Volume X. Redwood City, CA: Addison-Wesley 1991. Pages 603-629. 1991c.

....algorithms never assign credit directly to any component in the representation, we say they assign credit indirectly. Therefore, relative ranking of the population members as described above is sufficient feedback for learning in genetic algorithms. Koza s Genetic Programming Paradigm [Koza92a] [Koza92b] replaces the binary string representation for population members with a primitive language arranged into expression trees. The primitive language relies on a simple and uniform syntax to remove the possibility of generating an expression tree that is invalid with the recombination operators. ....

Koza, J. "Genetic Evolution and Co-Evolution of Computer Programs." In Artificial Life II, edited by C. Langton, C. Taylor, J. Farmer and S. Rasmussen. Reading, MA: Addison-Wesley Publishing Company, Inc., 1992.

....modeling [ Huston et al. 1988 ] they have been essential to artificial life. Emergent dynamics are often observed through the local interaction of individuals. This is the basis of cellular automata models as well as many of the early models in artificial life [ Langton, 1992, Fontana, 1992, Koza, 1992, Taylor et al. 1989, Travers, 1989, Reynolds, 1987, for example ] Recently, attention 1 It should be noted that definitions of virulence vary amongst these authors. For example, Herre (1993) operationalizes virulence as a reduction in lifetime reproductive success of the host, rather than ....

John R. Koza. Genetic evolution and co-evolution of computer programs. In Christopher G. Langton et al., editor, Artificial Life II, pages 603-- 630, Redwood, CA, 1992. Addison-Wesley.

....recombination of features. 6 Although one might think that computer programs are so brittle and epistatic that they could only be genetically bred in a few especially congenial problem domains, we have shown that computer programs can be genetically bred to solve a surprising variety of problems [Koza 1989, 1990, 1991a, 1991b, 1992] A videotape visualization of the application of genetic programming can be found in Koza and Rice (1992a) 4 The Subsumption Architecture The conventional approach to building control systems for autonomous mobile robots is to decompose the overall problem into a series of functional ....

Koza, John R. 1991a. Genetic evolution and co-evolution of computer programs. In Langton, Christopher, Taylor, Charles, Farmer, J. Doyne, and Rasmussen, Steen (editors). Artificial 22 Life II, SFI Studies in the Sciences of Complexity. Volume X. Redwood City, CA: Addison-Wesley 1991. 603-629. .

....We have recently shown that entire computer programs can be genetically bred to solve problems in a variety of different areas of artificial intelligence, machine learning, and symbolic processing. Specifically, this recently developed genetic programming paradigm has been successfully applied (Koza 1989, 1990, 1991) to example problems in several different areas, including . machine learning of functions (e.g. learning the Boolean 11 multiplexer function) planning (e.g. navigating an artificial ant along an irregular trail. developing a robotic action sequence that can stack an arbitrary initial ....

Koza, John R. Genetic Evolution and Co-Evolution of Computer Programs. In Farmer, Doyne., Langton, Christopher, Rasmussen, S., and Taylor, C. (editors) Artificial Life II, SFI Studies in the Sciences of Complexity. Volume XI. Redwood City, CA: Addison-Wesley. 1991.

....genotype. Much current work on evolving control systems for robots involves encoding the entire explicit control program in the genotype see the Artificial Neural Networks (ANN s) of much current Artificial Life work [Collins, 1992, Harvey et al. 1993] or the Genetic Programming (GP) of Koza [Koza, 1991]. Here, the genotype simply states what is good and what is bad about the world (in each agent s perhaps unadaptive opinion) and nothing else. It says nothing about how to make good things happen and how to avoid bad things agents have to learn this via their own individual, historical ....

Koza, John R. (1991), Genetic evolution and co-evolution of computer programs, in Christopher G.Langton et al., eds., Artificial Life II.

.... programs and time limits (unlike, e.g. Genetic Programming (GP) LS may also be of interest for researchers working in GP and related fields among the first papers on using GP like algorithms to evolve assembler like computer programs are (Cramer, 1985; Dickmanns et al. 1987) See also (Koza, 1992) for later work. ALS: single tasks versus multiple tasks. If we use the adaptive LS extension (ALS) for a single task as the one above (by repeatedly applying LS to the same problem and changing the underlying probability distribution in between successive calls according to section 4.2) then ....

Koza, J. R. (1992). Genetic evolution and co-evolution of computer programs. In Langton, C., Taylor, C., Farmer, J. D., and Rasmussen, S., editors, Artificial Life II, pages 313--324. Addison Wesley Publishing Company.

....to solve problems. It is essentially an extension of a Genetic Algorithm (GA) 4, 1, 3] in which the individuals are programs. Koza has shown convincingly that the GP paradigm is general and provides a single, unified approach to many seemingly different problems in an astonishing variety of areas [5 10]. We believe GP to constitute a significant evolutionary extension and enrichment of the GA research paradigm itself; it is not just another application of GA methodology to a new domain. One reason is that GP manipulates variable length solutions which are not encoded as bit strings. String based ....

Koza, J.R., Genetic Evolution and Co-Evolutions of Computer Programs, in Artificial Life II, Editors: Langton, Taylor, Doyne Farmer, Rasmussen. (p 603-629), Addison Wesley, Redwood City, 1991.

....areas [Koza 1992] including . emergent behavior (e.g. discovering a computer program which, when executed by all the ants in an ant colony, enables the ants to locate food, pick it up, carry it to the nest, and drop pheromones along the way so as to produce cooperative emergent behavior) [Koza 1991a] planning (e.g. navigating an artificial ant along an irregular trail) Koza 1990b] finding minimax strategies for games (e.g. differential pursuer evader games; discrete games in extensive form) by both evolution and co evolution [Koza 1991b] optimal control (e.g. centering a cart and ....

.... so as to produce cooperative emergent behavior) Koza 1991a] planning (e.g. navigating an artificial ant along an irregular trail) Koza 1990b] finding minimax strategies for games (e.g. differential pursuer evader games; discrete games in extensive form) by both evolution and co evolution [Koza 1991b] optimal control (e.g. centering a cart and balancing a broom in minimal time by applying a bang bang force to the cart) Koza and Keane 1990a, 1990b] machine learning of functions (e.g. learning the Boolean 11 multiplexer function) Koza 1991d] generation of random numbers (using ....

[Article contains additional citation context not shown here]

Koza, John R. Genetic evolution and coevolution of computer programs. In Langton, Christopher, Taylor, Charles, Farmer, J. Doyne, and Rasmussen, Steen (editors). Artificial Life II, SFI Studies in the Sciences of Complexity. Volume X. Redwood City, CA: Addison-Wesley 1991. 603-629. 1991a.

....of means (other than the radiation procedure reported here) by which premature convergence might be prevented; e.g. one would could segregate the string population into subpopulations with limited inter group breeding. One might even attempt, e.g. via the Genetic Programming approach of Koza [1991], to evolve not string solutions to particular Patterson functions, but RPP solving algorithms which would then be generally applicable. Similarly, there are no doubt numerous ways by which more powerful forms of Simulated Annealing might be applied to the RPP. AT the least, the SA procedure could ....

Koza, J. R. [1991]. "Genetic Evolution and Co-Evolution of Computer Programs." Langton et all, eds., Artificial Life II (see above, Belew reference), pp. 603-629.

....programming paradigm, populations of entire computer programs are genetically bred to solve problems. We have recently shown that entire computer programs can be genetically bred to solve problems in a variety of different areas of artificial intelligence, machine learning, and symbolic processing (Koza 1989, 1990a, 1991d) In this recently developed genetic programming paradigm, the individuals in the population are compositions of functions and terminals appropriate to the particular problem domain. The set of functions used typically includes arithmetic operations, mathematical functions, conditional ....

....to example problems in several different areas, including . machine learning of functions (e.g. learning the Boolean 11 multiplexer function) planning (e.g. navigating an artificial ant along an irregular trail, developing a robotic action sequence that can stack blocks in a specified order) (Koza 1990c) automatic programming (e.g. solving pairs of linear equations, solving quadratic equations for complex roots, and discovering trigonometric identities) optimal control (e.g. centering a cart and balancing a broom on a moving cart in minimal time by applying a bang bang force to the ....

[Article contains additional citation context not shown here]

Koza, J. R. (1991a). Genetic Evolution and Co-Evolution of Computer Programs. In Artificial Life II, SFI Studies in the Sciences of Complexity. (D. Farmer, C. Langton, S. Rasmussen and C.

.... plan for stacking blocks in to a desired order) 4, 6] emergent behavior (e.g. discovering a computer program for locating food, carrying food to the nest, and dropping pheromones, which, when executed by all the ants in an ant colony, produces interesting higher level emergent behavior) [10], machine learning of functions (e.g. learning the Boolean 11 multiplexer function) 11] automatic programming (e.g. solving pairs of linear equations, solving quadratic equations for complex roots, and discovering trigonometric identities) 4] discovering inverse kinematic equations (e.g. ....

J. R. Koza. Genetic evolution and co-evolution of computer programs. In Farmer, Doyne, Langton, Christopher, Rasmussen, S., and Taylor, C. (editors) Artificial Life II, SFI Studies in the Sciences of Complexity. Volume XI. AddisonWesley, Redwood City CA 1991.

No context found.

J. Koza. Genetic evolution and co-evolution of computer programs. In J. D. F. Christopher G. Langton, Charles Taylor and S. Rasmussen, editors, Artificial Life II. Addison Wesley Publishing Company, Reading Mass., 1992.

No context found.

J. R. Koza. Genetic evolution and co-evolution of computer programs. In C. Langton, C. Taylor, J. Farmer, and S. Rasmussen, editors, Artificial Life II, volume X of SFI Studies in the Sciences of Complexity, pages 603--629. Addison-Wesley, Redwood City, CA, 1992.

No context found.

John R. Koza. Genetic evolution and co-evolution of computer programs. In Christopher Taylor Charles Langton, J. Doyne Farmer, and Steen Rasmussen, editors, Artificial Life II, volume X of SFI Studies in the Sciences of Complexity, pages 603--629. Addison-Wesley, Santa Fe Institute, New Mexico, USA, February 1990 1991.

No context found.

Koza, J.R. (1992), "Genetic evolution and co-evolution of computer programs," Artificial Life II, pp. 313--324, Addison Wesley Publishing Company.

No context found.

Koza, John R. (1991), Genetic evolution and co-evolution of computer programs, in Christopher G.Langton et al., eds., Artificial Life II.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC