Composable Agents for Patient Flow Control | Preliminary Concepts

In this article we describe our research eeorts in coping with a trade-oo that can be often found in the control and optimization of todays business processes. Though centralized control may achieve better results in controlling the system behavior, there are usually social, technical and security constraints on applying centralized control. Distributed control on the other hand may cope with these constraints but also entails sub-optimal results and communicational overhead. Our concept of composable agents tries to allow a dynamic and uent transition between globalization and local-ization in business process control by adapting to the current real-world system structure. We are currently evaluating this concept in the framework of a patient ow control project at Charit e Berlin. Todays applications of information technology face at least two major aspects of business settings. The rst aspect is the partially or fully automated execution of complex business processes. This enforces the use of sophisticated methods, like monitoring, planning and scheduling. The second aspect is the control of the process execution itself. Since most real-world processes are distributed spatially and among individuals, the coordination and interaction between the execution entities is a question of substance to process management. The notion of an intelligent agent (Huhns & Singh 1998b; Jennings & Wooldridge 1998) is a recent concept that tries to incorporate the merits of sophisticated AI methods with strong communication and interaction capabilities. Intelligent agents have traditionally been applied to the control and optimization of industrial transport and production. In contrast to that, our research is more involved with human processes in the domain of administration and health care management. In the following, we will present the problem of distributed patient ow control which can be seen as a problem for collaborative problem solving in a business context. Then we will introduce the new concept of composable agents that is tailored to dynamic and adaptive reconnguration in such collaborative problem solving situations. Berlin and GMD FIRST are currently carrying out an industrial-size case study at the cardiological clinic of Charit e Berlin, Europe's biggest hospital. The cardiological clinic consists of ve wards with a capacity of altogether over 80 patients, four outpatients' facilities, in which diierent types of medical consulting are done in parallel, and eight diagnostic units, some of which with several subunits. The diagnostic units perform over 100 diagnostic examinations each day. These examinations are requested by the wards, the outpatients' department …