People

Paul Baguely

I need a title, Ph.D. Thesis (2005)

Research Project Outline

Global competition has correspondingly increased the required number of new cost models. These new cost models must be rapidly built, use less data, operate using expert opinion, and be understandable to a wide variety of product team members.

The cost model development process consists of data identification, data collection, and data analysis tasks. Fuzzy logic is considered as a new method to fulfil the new requirements in cost model development. Fuzzy logic can be built using data or words, and can be similarly understood by users.

The fuzzy logic methods of Mamdani, a subtractive clustering based algorithm, and the Adaptive Neuro-Fuzzy Inference System are chosen from a review of the fuzzy logic literature, to be compared with Multiple Linear Regression analysis. These methods are subsequently tested on a non-linear and linear function form, comparable to cost model forms found in the literature. Different configurations of the fuzzy logic methods are tested and built efficiently through the Taguchi methodology. Dependant on the error measures of Average Percentage Error and Average Absolute Error, each method is placed in a favourable light through a particular chosen configuration by the Taguchi methodology. It is found that all methods were capable in at least one instance of estimating a cost model function to a definitive level, a level suitable for commercial quotations. In addition there was also a possible range of errors dependant on the fuzzy logic model configuration.

The numerical results are used to assist in the development of a proposed advisory process for cost engineers using fuzzy logic. The process levers case studies from the literature, the numerical results from the experiments, and knowledge obtained through the research that is captured within the thesis.

After leaving De Montfort University Paul worked at Durham University from July 2003 to February 2006. At Durham Paul worked on the Digital Enterprise Technology (DET) project in which he developed the specification for a DET Cost Engineering toolbox. Paul was involved with several industrial organisations developing concepts in Cost Management, lean and agile manufacturing and life cycle costing. In March 2006 Paul moved to the Decision Engineering Centre at Cranfield University where he is a Research Fellow in Cost Engineering. Paul develops training courses in Cost Engineering, proposals for research projects and supervises a student at BAE Systems working on uncertainty at the bidding stage.