No abstract

No abstract

Abstract. In evolutionary computation many different representations ("genomes") have been suggested as the underlying data structures, upon which the genetic operators act. Among the most prominent examples are the evolution of binary strings, real-valued vectors, permutations, finite automata, and parse trees. In this paper the use of place-transition nets, a low-level Petri net (PN) class [1,2], as the structures that undergo evolution is examined. We call this approach "Petri Net Evolution" (PNE). Structurally, Petri nets can be considered as specialized bipartite graphs. In their extended version (adding inhibitor arcs) PNs are as powerful as Turing machines. PNE is therefore a form of Genetic Programming (GP). Preliminary results obtained by evolving variablesize place-transition nets show the success of this approach when applied to the problem areas of boolean function learning and classification. OverviewPNs are a nice formalisms that allow the complex modeling of numerous real world problems. Their strength is the "built in" concurrency mechanism, which allows a formal description of systems containing parallel processes. For this study small sample problems have been used to test the feasibility of PNE. The PNs evolved belong to a class of low level PNs called place-transition nets.This work has been inspired by the generalization of the recombination operator from trees to bipartite graphs. Since recombination seems to be an important genetic operator in PNE, this approach follows the GA stream within EC.The common feature of traditional GP and PNE is the variable length genotype. This is in contrast to the standard GA which employs fixed-length strings. In contrast to traditional tree-based GP, there is no need to specify a function set explicitly for PNE. Functions are implicitly built into the semantics of the chosen PN model. Because the place-transition PN model used is integer-based our focus is on discrete problems rather than continuous ones.PNE evolves a population of PNs in an evolution-directed search of the space of possible PNs for ones, which, when executed, will exhibit high fitness. To the best of our knowledge the idea of evolving a population of individuals which are variable-sized PNs does not appear in previous literature.

No abstract

No abstract