Intelligence engineering framework

A number of advanced algorithms and mostly software-based technologies have been developed in recent decades in order to solve problems in complex technical systems. Examples that have been actively studied include machine learning, artificial intelligence, pattern recognition, neural networks, fuzzy logic, statistical methods, operation analysis and, most recently, sensor networks. The problem is that these techniques require considerable knowledge to be applied correctly, necessitating participation by skilled professionals. This makes applications expensive to design and maintain. There is therefore a common need for better engineering methods and tools. This paper describes start of the development of a systematic engineering discipline for algorithmic and knowledge-intensive intelligent systems and services. The rationale behind this idea is that advanced technologies and algorithms cannot be economically feasible unless standardised design practices, tools and system components are available. The focus of the research on the early stages of design gave rise to two issues of design reuse: 1) how to model the application's needs for intelligence and the features of potential solutions stored in solution libraries, and 2) how to help the designer search the libraries for solutions that provide the best match with the application needs.