Generic business process modelling framework for quantitative evaluation

Abstract

Business processes are the backbone of organisations used to automate and increase the efficiency and effectiveness of their services and prod- ucts. The rapid growth of the Internet and other Web based technologies has sparked competition between organisations in attempting to provide a faster, cheaper and smarter environment for customers. In response to these requirements, organisations are examining how their business processes may be evaluated so as to improve business performance. This thesis proposes a generic framework to expand the applicability of various quantitative evaluation to a large class of business processes. The framework introduces a novel engineering methodology that defines a modelling formalism to represent business processes that can be solved for a set of performance and optimisation algorithms. The methodology allows various types of algorithms used in model-based business pro- cess improvement and optimisation to be plugged in a single modelling formalism. As a part of the framework, a generic modelling formalism (MWF-wR) is developed to represent business processes so as to allow quantitative evaluation and to select the parameters for the associated performance evaluation and optimisation. The generic framework is designed and implemented by developing soft- ware support tools using Java as object oriented programming language combining three main modules: (i) a business process specification mod- ule to define the components of the business process model, (ii) a stochas- tic Petri net module to map the business process model to a stochastic Petri net, and (iii) an algorithms module to solve the models for various performance optimisation objectives. Furthermore, a literature survey of different aspects of business processes including modelling and analy- sis techniques provides an overview of the current state of research and highlights gaps in business process modelling and performance analy- sis. Finally, experiments are introduced to investigate the validity of the presented approach.