M. Sipper, #Non-uniform cellular automata: Evolution in rule space and formation of complex structures," in Arti#cial Life IV, eds. R. A. Brooks and P. Maes, #The MIT Press, Cambridge, Massachusetts, 1994#, pp. 394#399.  Home/Search   Document Details and Download   Summary   Related Articles   Check

.... of universal constructors based on von Neumann s self reproducing automaton, studied in 50s 70s (von Neumann 1966; Codd 1968; Vit anyi 1973; Pesavento 1995) 2) Search for a minimal system capable of non trivial self replication, studied in 80s 90s (Langton 1984; Byl 1989; Reggia et al. 1993; Sipper 1994; Morita Imai 1996b) 1 See also (Sayama 1998a, Chap.3) and (Sipper 1998) 3) Addition of other computational capabilities to selfreplicators, studied in 90s present (Tempesti 1995; Perrier, Sipper, Zahnd 1996; Chou Reggia 1998) 4) Realization of emergence and evolution of ....

Sipper, M. 1994. Non-uniform cellular automata: Evolution in rule space and formation of complex structures.

....system in which global information processing appears as a result of the interactions among many components, each of which is a system which in turn exhibits an ability for global computation at a different level of self organisation. The model, derived from the non uniform cellular automata model [1 4], is inspired by the idea of co evolving to a higher degree of specialisation with macrostructures in a single computational ecosystem. For other relevant work, we refer to the ecological models studied in [5 7] Cellular automata (CAs) are discrete dynamical systems of simple locally connected ....

M. Sipper. Non-uniform cellular automata: Evolution in rule space and formation of complex structures. In R. A. Brooks and P. Maes, editors, Artificial Life IV, pages 394--399, Cambridge, MA, 1994. MIT Press.

....so on (Fig. 3.6) 42] Besides the modification of Langton s loop, there were several interesting challenges to realize self reproduction in other forms. Sipper showed that it is much easier to implement many lifelike phenomena by using non uniform non deterministic CA than by using traditional CA[49]. Morita et al. succeeded in creating self reproducing structures with a shape encoding mechanism with which every structure (worms as well as loops) can reproduce itself in a non trivial way (Fig. 3.7) 32, 33] 3.4.3 Applying to practical problems The application of self replicating automata ....

Sipper, M. : Non-uniform cellular automata: Evolution in rule space and formation of complex structures, in Brooks, R. and Maes, P. eds., Artificial Life IV: Proceedings of the Fourth International Workshop on the Synthesis and Simulation of Living Systems, pp. 394--399, Cambridge, Massachusetts (1994), MIT Press.

....implemented and experimented. The model is derived from two classical formal machines: Turing Machines [Turing36 ] and Cellular Automata [Ulam50] The first idea is to extend CAs into a machine with some built in features supporting movement. While adding movement to CAs has also been explored in [GoelThomp88, Sipper94, Eckart95], one of the main differences to those works is the explicit integration in the model of energetical constraints. Important features issued from genetics and neural networks are also integrated in the model. The proposed model is intended to be a theoretically well founded and simple tool leading ....

....this operation is successful, the energy of the mobile cell is increased by a value that only depends on the eaten symbol. dynamics of the environment: We relax some constraints on the environment by allowing also nondeterministic and non uniform (or inhomogeneous) CAs. As already pointed in [Sipper94], this preserves most of the essential features of the original model of CAs: locality of interactions, massive parallelism and simplicity of the cells (FSMs) The reason is that that property is not absolutely necessary in our context: CAs are used here to model environments and not the universe. ....

M. Sipper, "Non-Uniform Cellular Automata: Evolution in Rule Space and...," Artificial Life IV, MIT Press, 1994.

No context found.

M. Sipper, #Non-uniform cellular automata: Evolution in rule space and formation of complex structures," in Arti#cial Life IV, eds. R. A. Brooks and P. Maes, #The MIT Press, Cambridge, Massachusetts, 1994#, pp. 394#399.

....by [ Mitchell et al. 1993, Mitchell et al. 1994b, Das et al. 1994 ] who demonstrated that high performance CA rules can be evolved using genetic algorithms. We have investigated an extension of the CA model termed non uniform cellular automata, in which cellular rules need not be identical [ Sipper, 1994, Sipper, 1995b, Sipper, 1995a ] Employing this model we found that high performance can be attained for the density task by means of co evolution [ Sipper, 1996 ] 1 . Non uniform CAs have also been investigated by [ Vichniac et al. 1986, Hartman and Vichniac, 1986 ] As noted by Mitchell et ....

M. Sipper. Non-uniform cellular automata: Evolution in rule space and formation of complex structures. In R. A. Brooks and P. Maes, editors, Artificial Life IV, pages 394--399, Cambridge, Massachusetts, 1994. The MIT Press.

....who discuss a onedimensional CA in which a cell probabilistically selects one of two rules, at each time step. They showed that complex patterns appear, characteristic of class IV behavior 2 (see also [Hartman and Vichniac, 1986] The issue of evolving non uniform CAs was discussed by us in [Sipper, 1994, Sipper, 1995b] where complex phenomena were observed, though we did not evolve CAs to perform a particular computational task. In [Sipper, 1995a] we examined the issue of universal computation in non uniform CAs (see Section 5) A prime motivation for studying non uniform CAs stems from the ....

M. Sipper. Non-uniform cellular automata: Evolution in rule space and formation of complex structures. In R. A. Brooks and P. Maes, editors, Artificial Life IV, pages 394--399, Cambridge, Massachusetts, 1994. The MIT Press.

....appearance on the artificial evolution scene [ Sanchez and Tomassini, 1996 ] A prime motivation of the work reported in this paper stems from our desire to attain more realistic systems that are amenable to implementation as evolware. 2 Previous work We had first studied non uniform CAs in [ Sipper, 1994, Sipper, 1995b ] and demonstrated in [ Sipper, 1995a, Sipper, 1996b ] that universal computation can be attained in such CAs. The evolution of non uniform CAs was undertaken in [ Sipper, 1996a, Sipper, 1996c, Sipper and Ruppin, 1996a, Sipper and Ruppin, 1996b, Sipper and Tomassini, 1996a, Sipper ....

M. Sipper. Non-uniform cellular automata: Evolution in rule space and formation of complex structures. In R. A. Brooks and P. Maes, editors, Artificial Life IV, pages 394--399, Cambridge, Massachusetts, 1994. The MIT Press.

....Marchal et al. 1994] Their intent is to create an architecture which is complex enough for (quasi) universal computation yet simple enough for physical implementation; among the properties demonstrated by this group is a form of multi cellular reproduction. Finally, we mention the work of [Sipper, 1994, Sipper, 1995] in which a non uniform, CA derived model is presented, and used to implement various systems, among them a self reproducing, multi cellular loop. 3 Self inspection based methods were first introduced by [ Laing, 1975, Laing, 1976, Laing, 1977 ] using a different model than that ....

M. Sipper. Non-uniform cellular automata: Evolution in rule space and formation of complex structures. In R. A. Brooks and P. Maes, editors, Artificial Life IV, pages 394--399, Cambridge, Massachusetts, 1994. The MIT Press.

....of our primary goals is to increase the capacity for AL modeling while preserving the essential features of the original CA, namely massive parallelism, locality of cellular interactions and simplicity of cells. Thus we attain a model that is close to CA with regards to generality and simplicity [24, 25]. In this paper we consider non uniform CA, i.e. with non uniform local interaction rules. Such automata function in the same way as uniform ones, the only difference being in the cellular rules which need not be identical for all cells. This type of model has been investigated by others. 8] ....

.... where an automaton is trained to solve some problem (e.g. pole balancing) As noted above, our motivation is different than these works: we wish to study Artificial Life phenomena such as evolution, emergence and multi cellularity in a simple, general model which operates at a higher level than CA [24, 25]. The focal point of this paper is the issue of universal computation in two dimensional CA, namely the construction of machines, embedded in cellular space, whose computational power is equivalent to that of a universal Turing machine [10] The first such machine was described by von Neumann, who ....

[Article contains additional citation context not shown here]

M. Sipper. Non-uniform cellular automata: Evolution in rule space and formation of complex structures. In R. A. Brooks and P. Maes, editors, Artificial Life IV, pages 394--399, Cambridge, Massachusetts, 1994. The MIT Press.

....(1) massive parallelism, 2) locality of cellular interactions and (3) simplicity of cells (finite state machines) The basic model is detailed in Section 2 and the evolutionary aspect is presented in Section 4.1. We delineate below the three basic features by which it differs from the CA model [Sipper, 1994]: 1. Whereas the CA model consists of uniform cells, each containing the same rule, we consider the non uniform case where different cells may contain different rules. 2. The rules are slightly more complex than CA rules. 3. Evolution takes place not only in state space as in the CA model, but ....

....in depth studies of the evolutionary process. This is accomplished in our case by observing not only phenotypic effects (i.e. evolution of cell states as a function of time) but also fitness, operability, energy and the genescape. Our approach may be viewed as a non uniform CA with enhanced rules [Sipper, 1994] and as such it is related to other works. Garzon, 1990] presents two generalizations of cellular automata, namely discrete neural networks and automata networks. These are compared to the original model from a computational point of view which considers the classes of problems such models can ....

M. Sipper. Non-uniform cellular automata: Evolution in rule space and formation of complex structures. In R. A. Brooks and P. Maes, editors, Artificial Life IV, pages 394--399, Cambridge, Massachusetts, 1994. The MIT Press.

No context found.

M. Sipper. Non-uniform cellular automata: Evolution in rule space and formation of complex structures. In R. A. Brooks and P. Maes, editors, Artificial Life IV, pages 394--399, Cambridge, Massachusetts, 1994. The MIT Press.

No context found.

M. Sipper. Non-uniform cellular automata: Evolution in rule space and formation of complex structures. In R. A. Brooks and P. Maes, editors, Artificial Life IV, pages 394--399, Cambridge, Massachusetts, 1994. The MIT Press.

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