Knowledge based improvement : simulation and artificial intelligence for understanding and improving decision making in an operations system

摘要

The thesis investigates the possibility of using simulation for understanding and improving the design of decision making in a real context. The approach is based on the problem of representing decision making behaviour in Discrete Event Simulation. An investigation of existing techniques led to the design of a methodology known as Knowledge Based Improvement (KBI). The KBI covers the key stages of the process of using simulation for understanding and improving the design of decision making. Using a research strategy that involves a case study in Ford, the research tests each stage of KBI. The thesis explains how simulation can be used for understanding real decision making problems and for collecting the data required for modelling individual decision making strategies. The thesis demonstrates the possibility of a simulation based knowledge elicitation in a real context and it investigates the practical difficulties involved in this process. The research tests the process of understanding decision making policies by modelling specific decision makers using Artificial Intelligence. It tests the use of simulation for assessing the decision making strategies and it shows that simulation can be used for identifying efficient strategies and for improving the design of decision making practices. The thesis reports the degree of success of the approach in relation to the data that were collected and it describes the validation checks that were undertaken. In addition, it reports the lessons learned from the application of the KBI methodology, the overall success of the approach and the main limitations that were identified during the implementation.