Model Management for Cybernetic Decision Support Systems

Managers being human have limited information processing capacity and are subject to judgmental biases, inferential shortcomings, and ignorance of the rules for optimal information processing and decision-making. Decision aids in the form of human computer information processing systems such as decision support systems, are sometimes employed to assist and support human managers in various tasks. These decision aids unfortunately do not fulfill the requirements and expectations of managers in arriving at the desired solution. This is due to the fact that the tools provided are not suitable for the managers, as they do not put sufficient emphasis on the human aspects of decision-making. Most recent models for decision support systems presented by researchers in model management are based on Operational Research or Artificial Intelligence, and are not adequate. They are based on a simplistic question-answer environment which does not reflect the real-life situation. They also do not put enough emphasis on the intelligent, deciding and reasoning side of the model. Furthermore, the models have little capacity to learn and adapt to new environments and needs. Thus, the proposal in this thesis is for a new system called Model Manager System (MOMS) that incorporates Artificial Intelligence and a Cybernetic Approach with the actual Decision-Making Environment. In order to design the proposed decision model system, various areas are explored - such as the decision-making process, managerial behaviour, human information system, and available decision aids - where various elements related to human decision making are considered. As Cybernetic tools are used in designing the model, human aspects are emphasised greatly, especially in the Information Processing techniques such as intelligence, control, coordination, monitoring, and implementation. The designed system tries to mimic Human Information Processing wherever possible.

Preview PDF 549611_FSKTM_2004_9.pdf Download (105kB)

Actions (login required)

View Item