Keywords: soft computing, fuzzy logic, genetic algorithms, neural networks, building guidance, prescriptive vs. performance-based norms.

Soft computing offers novel methodologies that reflect the ability of the human mind to process speedily and store large amounts of information, to exploit the tolerance for imprecision and uncertainty, to achieve at low cost tractability and robustness. This paper shows the need for engineering applications to be considered alongside building standards or norms integrated within a suitable assurance and regulatory framework so as to recognise the paradigm shift from hard computing requirements of precisely defined analytical models and vast amounts of computation time to knowledge- and data-driven computing. Despite extensive study and application for scientific research and engineering computing since 1980s many questions remain unanswered. A review of the literature identified important issues in soft computing engineering applications and their inter-relationship with building standards or norms, certification and regulatory frameworks. A major aim was increasing our understanding of the effect of the various types of new technologies on building norms, assurance and the regulatory framework.

Delineating meanings and scope alongside benefits and limits of the new knowledge from soft computing engineering applications relative to building standards were an essential component and the first step of this survey. Existing guidance or standards, whether prescriptive or performance-based, not only underpin building safety thresholds and define risk management strategies but also provide the baseline measure of quality, a useful basis for any design quality improvements. Prescriptive requirements spell out exactly how something is to be done while performance requirements merely outline what the required level of performance is, and leave it up to the designers as to how this is achieved. Historically, prescriptive requirements are very reactive and depend largely on a problem occurring followed by changes to the industry guidance or standards to ensure that the problem never happens again.

The review of the literature established the need for publications on the impact of soft computing on building norms. Much literature was about historical development of expert systems, fuzzy logic, neural networks and evolutionary algorithms, explaining theory, specific algorithms and laboratory applications. No publications focused on consequences of applying these novel technologies on building norms. One conclusion is for several essential requirements. A regulatory framework that keeps pace with the rapid application of soft computing in various fields of engineering is required to provide adequate safeguards against breaches of safety or poor design solutions. Similarly the costs of conducting the essential development work, testing and validating industry standards or norms need to be allocated between public and private sectors. Identifying measures to support engineering teams in capturing and sharing evidence, lessons learnt from projects applying soft computing and clarifying the role of assurance are required. All these are important areas for further research studies.

purchase the full-text of this paper (price £20)

go to the previous paper
go to the next paper
return to the table of contents
return to the book description
purchase this book (price £70 +P&P)