University of the West of England Home Page - Peter Hale Home Page - SEEDS Site Map SEEDS Page - Text Only Site Map

Systems Engineering Estimation and Decision Support (SEEDS)

The SEEDS team is part of the Aerospace Manufacturing Research Centre (AMRC) - http://www.uwe.ac.uk/cems/research/centres/amrc.html. Our expertise is in working with aerospace organisations' to apply techniques in managing, categorising and visualising information to support costing of complex products mainly within the aerospace sector.

The tools and techniques we provide can be applicable to organisations outside this field. We are experienced at liasing with industrial partners to ensure our projects meet their needs and so have a successful outcome. Our main area of research is in software techniques for analysing and reporting on the costs involved in design and manufacture of products, and integrating such tools into the infrastructure that exists in an organisation.

We are involved in cost modelling solutions, and have experience of working with aerospace organisations. We have been working with Rolls-Royce aerospace and Airbus.

Research

Research projects have involved providing means to visualise cost information generated by estimation packages, a task severely missing at the present day. Visualisation of the cost information will permit analysing the cost information in depth so as to locate not only over cost areas, but also areas of cost which are problematic for other reasons (e.g. uncertainty), or which present the best cost reduction opportunities. Prior to visualising information, it was necessary to assess cost estimation methodologies so as to evaluate the usefulness of their output for visualisation.

Parametric Cost Estimation (PCE) techniques were first evaluated. It was determined that despite their requirements of a low volume of input data, making them suitable at stages where little is known about a product, their output did not provide enough details to benefit from a visualisation tool, unless a large number of such models are used to increase the data volume (which would be time consuming and costly).

Research was then carried out to study an alternative to PCE, namely Generative Cost Estimation (GCE). GCE do generate an ideal input to hypothetical cost visualisation tools in that an extensive hierarchical cost information structure can be produced, containing simultaneously numerous data fields (e.g. time and uncertainty in addition to a cost figure), at a detailed level (e.g. in the form of a complete Bill Of Materials, or as a process sequence).

Using CGE it is possible to generate costs from fundamental equations, process models and accounting rules. Hence the cost estimate can be compared with parametric models and any differences fully rationalised and understood.

A visualisation tool was created. It was vital to ensure that this tool could read data from a wide number of formats to make it compatible with numerous cost estimation packages used by our partners (e.g. Microsoft Excel, Microsoft Access, DecisionPro, and generic formats such as text, Comma Separated Values and eXtensible Markup Language XML). The tool contains a combination of hypertextual and exploratory techniques allowing for both knowledge discovery and data analysis in one package. Graphical representations were used to take advantage of human cognition, thus increasing knowledge discovery efficiency, mainly thanks to a graphical hierarchies representation tool named Cost Map. The efforts invested into easing the migration of knowledge from depositories or generation packages to the visualisation tool were repeated to ensure that the Cost Map itself could be distributed easily to a wide range of users, independently from their location, hardware and software configurations. This was possible by implementing exporting facilities in platform independent formats such as raster pictures as well as vector graphics in the Scalable Vector Graphics (SVG) format. These outputs make the Cost Map usable in documents produced by packages such as the Microsoft Office suite, as well as interactive web based pages. Lighter SVG versions were also produced for compatibility with mobile devices (e.g. Pocket PCs and mobile phones offering a particular SVG implementation viewer).

This approach makes it possible to display a holistic view of costs for a complex product. The use of a cost map permits designers to quickly identify priority areas for a complex design, which may have many levels of assembly and many thousands of parts. Hence, for example, the designer will be able to identify areas of high cost uncertainty (therefore potentially high cost risk) or areas of high relative cost.

Research has been undertaken into the ability to analyse cost with stochastic input variables. Hence enabling the cost model to allow designers to describe the current level of uncertainty of the product attributes (and possibly process uncertainty for new or poorly understood processes). This allows a cost estimate to be computed as a distribution. Further work has been undertaken into allowing wide access to cost models by the use of web-based tools that run through standard web browsers.

We have developed and used techniques to enable decision support during product development, whilst minimising dependence on specialist software and detailed programming effort. The basis of this is an Ontology that can be visualised and edited in tree form. We are using the open standard Stanford University Ontology tool Protégé. This Ontology can be translated into a Decision Support tool called DecisionPro. Software we have created using DecisionPro allows calculations of the cost of a design to be made, and provides a colour-coded representation of the product tree. It is then possible to output this tree in the form of web pages, interactive diagrams and code in programming languages such as the Java based costing tool Cost Estimator. This enables people who are computer literate but are not software developers to produce decision support and cost models. It is possible to search the information both in Protégé and on the Web as it is represented using searchable semantic web languages.