Engineering Design


The engineering of methods and tools to support the automation of engineering design is a particular interest of factasia.
Engineering design is a process in which an understanding of the physical world, sometimes embodied in quantitative (i.e. mathematical) scientific models, is used to devise a recipe for constructing physical entities (e.g. machines, buildings) or information structures (e.g. software) which satisfy some real human need. Factasia is concerned with the modelling methods and tools which are made available to support engineering design. We look forward to a future in which formal descriptions of required behaviour and of proposed structure enable automatic simulation for the evaluation of requirement and design, and in which these same formal descriptions provide the basis for increasing levels of automation of the design process.
Computer Aided Design
The use of computers to assist in the engineering design process will gradually become pervasive. Initially simulation is often accomplished by algorithmic techniques, where the required behaviour is simply programmed into a simulator.
Mathematical Modelling
The Mechanisation of Mathematics (through Intelligent Logic) will pave the way for Mathematical modelling in which declarative descriptions supplant algorithmic descriptions. These declarative descriptions are easier to produce and permit not only simulation but other more advance logical analysis. Such descriptions will be application independent and will permit better computer support for the design process. They also permit formal specifications to be provided prior to design and automation of
Automated Design
Gradually human designer will focus more on providing specifications and evaluating designs produced to an increasing extent by the Intelligent Logic of The Global SuperBrain from formal requirements specifications. Intelligent interfaces will eliminate much of the difficulty in preparing these specifications, evaluating and fine tuning design proposals.
Machine Intelligence
The ultimate goal is the automation of design through the mechanisation problem solving using mathematical models. The idea is to start with people providing the intelligence and computers providing good but not really "intelligent" support, mainly through mathematical modelling. Then gradually we automate larger parts of the process and increase the problem solving capabilities of the software until there is a real collaboration between human designer and machine in creating a design which meets the requirements. Ultimately it might be that the computers do all the design and human engineers are concerned only with working out the requirements, setting design standards, and further developing the software.
Analytic Engineering
The adaptation of engineering methods to exploit the mechanisation of deductive (and with it mathematics) we call analytic engineering. The idea is that just as we now use machines for computation, we will in the future use them for design.
Formal Models for Engineering
In order for engineering design to benefit from the automation of deduction the scientific basis for engineering design must be formalised. The informal quantitative laws which engineers use for predicting the behaviour of some proposed engineering design must be replaced by formal models permitting the behaviour of the design to be formally derived.
Formal Requirements
It is also necessary for the statement of the problem to be solved to be given with formal precision. This stage in the engineering process would therefore deliver a formal result, but would not be a logical activity. It involves the translation of an informal understanding of the requirements into formal notations. How much of this formality is visible to the people involved in the process is likely to vary greatly.


UP HOME © RBJ created 1998/08/06 modified 1998/08/06