Article Preview
Top1. Introduction
A business process is defined as consisting of a set of activities, performed by their relevant roles or collaborators, to intentionally achieve the common business goals (Smith & Fingar, 2003). Business processes are the core assets of any enterprise since they generate revenue and often represent a significant proportion of costs. A recent study (Hill, Cantara, Deitert, & Kerremans, 2007) has shown that the business process management (BPM) software market reached nearly $1.7 billion in total software revenue in 2006 and this number continues to grow. The flexibility and reactivity in process execution through IT-systems have significant impact on the success of an organisation’s business operations, especially in the current constantly changing business environment. Existing BPM systems, which require a priori representation of a business process and all potential deviations from that process, however, do not provide adequate support to achieve these requirements in a satisfactory way (Burmeister, Arnold, Copaciu, & Rimassa, 2008; Fingar, 2008).
A software agent (Wooldridge, 2002) is a piece of software which is situated in an environment, autonomous (i.e. acts on its own), social (interacts with other similar entities), and being reactive (responding to changes in its environment) and/or being proactive (working to achieve its goals). Multi-Agent systems (MAS) provide powerful and flexible execution platform for business processes. On such a platform, the executing process is able to pursue persistent goals over time despite previously failed attempts due to the availability of multiple ways of dealing with a given goal, as in the Beliefs-Desires-Intentions (BDI) platform (Rao & Georgeff, 1995). The agent view also provides an intuitive, well-suited level of abstraction for modelling and implementing process execution. For instance, a process participant can be explicitly represented as an agent at the execution level, which reflects more accurately the structure of an organisation. Existing process execution platforms such as the Business Process Execution Language1 (BPEL) do not capture such structural information at the business level (Endert, Küster, Hirsch, & Albayrak, 2007).
Since the late 1980s, the field of agent technology has attracted a substantial amount of interest from researchers (Jennings, Sycara, & Wooldridge, 1998; Luck, McBurney, Shehory, & Willmott, 2005). In particular, there have been a number of different agent theories, architectures, and languages proposed in the literature. Agent technology, however, still faces many challenges in being adopted by the industry despite its popularity and attractiveness as a research area (Weyns, Parunak, & Shehory, 2009). Therefore, closing the gap between the business community, BPM in particular, and agent technology can bring substantial benefits to both sides: agents gaining better industry traction whilst BPM having a powerful solution to deal with its current challenges (Ghose, 2009).