DECOMPOSING SIMULATION NETS BY TOKEN COMMUNICATION
A PHP Error was encountered
Message: Undefined index: id
Line Number: 218
The usual way to handle larger problems is to divide them into subproblems. Traditionally this is a weak point in Petri nets. Introduction of hierarchies in the net is one solution to the problem. An alternative way to decompose Petri nets is to model every subproblem as an independent net. These nets are linked together by communication between programmable object oriented tokens. INTRODUCTION AND BACKGROUND Simulation Nets (SNs) are Petri Nets extended for conven- ient modelling of discrete event simulation problems (Peterson 1981, Reisig 1985, Törn 1981, Törn 1991). Two tools have been developed for execution of SN models (Gustavson and Törn 1994a). Object Oriented (OO) Tokens, for increasing the modelling power of SNs, have been introduced in (Gustavson and Törn 1994b). Tokens are natural candidates to carry even complex code and the step from coloured tokens to object oriented tokens is very straightforward. If a token has identity, data and in particu- lar methods it can remember its history, make decisions based on knowledge and can learn by experience i.e. the basis for an intelligent choice is established. In order to extend the knowledge base and decision making of tokens we let them communicate with each other. This is the most significant difference between our token model and other similar models (Javor 1994, Lakos 1994). These programmable tokens with intercommunication facilities are very powerful in modelling. By their proper- ties they become the dynamic part of the otherwise global and static control structure of the net. They can be used for adding continuous and fuzzy properties to the net. In the following we will focus on another quite interesting con- sequence, namely modularisation.