Agent-enabled distributed decision support system generators : a framework and implementaiton

Abstract

Most traditional and modem decision support systems (DSSs) are built for efficiency but they fail to address flexibility in support for decision making. They provide insufficient resources needed in the decision making process, and inadequate support for decision makers to manage the processes for solving complex problems. Their ability to adapt to changes in user requirements and to provide support for the lifecycle of DSSs is typically constrained. The above situations worsen when resources are distributed across different locations, decision making processes are required to be integrated at different points in time, and when collaboration is needed among decision makers. As technology progresses, there is an increasing opportunity for addressing these deficiencies. This research builds frameworks and architectures for designing flexible DSSs to address problems and issues identified in the DSS literature. To identify the problems and issues, various decision making models are synthesised and DSSs are reviewed. DSS frameworks and architectures are then explored to address the identified problems and issues. A multi-methodological approach is adopted to build theories (i.e. the frameworks); these theories are realised by the architectures and implemented by prototyping, and the prototype is evaluated to improve the developed theories. These research processes are iterated until valid frameworks and architectures are established. To achieve the above goals, we propose a decision making process lifecycle framework. We leverage the power of the component independence and modelling concepts for designing DSS frameworks to support this process lifecycle. We develop DSS architectures to realise the frameworks with object oriented and agent paradigms. We implement a prototypical Agent-enabled Distributed Decision Support System Generator (ADDSSG) with software agent technology. This implementation is used as a proof of concept for the proposed decision making process model, frameworks and architectures. The DDSSG frameworks and ADDSSG architectures are domain independent. They have been evaluated using various decision making models. They were also evaluated against representative frameworks, architectures and implementations selected from the literature review. The ADDSSG implementation has been evaluated by business scenarios found in practice. It was also evaluated by system and domain experts. The evaluation results indicate that the proposed decision making process model, frameworks and architectures are valid and the implementation is valuable for the design of flexible DSSs.