Stochastic information technology modelling for business processes

Abstract

Business Processes (BP) and Information Technology (IT) are two areas that work very closely in helping organisations to keep or retain competitive advantage. Therefore, design in these areas should consider the advantages provided by, and the limitations that each of these domains imposes on each other. BP design tries to ensure that IT specifications are considered during the design of BP. Similarly, Information Systems (IS) design attempts to capture organisational needs, known as IS functional and Non-Functional Requirements (NFR), in order to meet the organisational goals. Despite this, BP and IT modelling techniques barely depict the way IT may affect BP performance or vice versa. For example, Business Process Simulation (BPS) is one of the modelling techniques that has been increasingly used to support process design. The performance measurements obtained from BPS models, though, are obtained considering only organisational issues, and thus cannot be used to assess the impact that IT may have on process performance. Similarly, IT modelling techniques do not provide IS performance measurements, and hence cannot depict the way IS may improve BP performance. The relationship between BP and IT can be alternatively described in terms of the relationships between BP, IS and Computer Networks (CN). By looking at the parameters that govern these relationships a simulation framework was developed, namely ASSESS-IT, that develops simulation models that provide performance measurements of BP, IS and CN, and thus can reflect the impact that IT (IS and CN) may have on BP performance. This research uses a case study to test the proposed framework (theory testing), to understand the way BP, IS, and CN domains interact (discovery), and to propose alternative theories to solve the problems found (theory building). The experimentation with the ASSESS-IT framework suggests that in order to portray the impact that IT may have on BP, analysts in these domains should first identify those performance specifications that describe how well the IS delivers its functionality (also known as non-functional requirements). It was found that when the IS does not depend on determined response time, the relationships between BP, IS and CN can be assessed using only the relationship between BP and IS. An alternative simulation framework, namely BPISS, is proposed to produce BPS models that provide performance measurements of BP and IS. Thus, BP and IT analysts can investigate the impact that a given IS design may have on BP performance, and identify a better BP and IS solution.