Modelling and optimizing processes of workflow nets by means of evolutionary techniques

Abstract

The thesis deals with the method of Analytical Programming, which is used for symbolic regression of Workflow Nets. These nets can be used to model various processes, which are demonstrated on several examples. The work will introduce the reader to the theory of Petri nets, Workflow nets, bio-inspired algorithms and the basics of many other methods. In the first part of the thesis, the state of the art is presented. The aims of this thesis are derived from this state of the art. The methods used and the design of the experiment are described below. An important part of the work is the verification of the generated nets quality, therefore a part of the thesis deals with methods for nets analysis. Methods for net analysis were tested on a generated dataset that contained nets of different sizes. The result of the experiment is presented and discussed. The functionality of tool for nets synthesis has been tested in several experiments, three of which are described in this work. These experiments were selected primarily because they demonstrate the solution of real life problems. Generated networks meet the requirements of well-structured networks, while providing a satisfactory solution to the task. At the end of the work, a discussion of the results and a proposal for future continuation of the area are laid out.